<< 25-08-2020 >>

---

00:06:18	*	leorize quit (Quit: WeeChat 2.8)
00:30:57	FromDiscord	<Varriount> disruptek: Cream of bacon sounds... suspicious.
00:31:58	FromDiscord	<Varriount> I guess you might make such a thing by mixing cream and bacon, and cooking them together?
00:34:31	FromDiscord	<Rosen> found a recipe for cream of bacon soup and it actually looks great
00:34:32	FromDiscord	<Rosen> at least to me
00:34:40	disruptek	i'll let you know how it comes out in a few minutes.
00:37:16	*	arecacea1 quit (Remote host closed the connection)
00:38:02	*	arecacea1 joined #nim
00:54:18	disruptek	better than expected and worse than i'd hoped.
01:09:32	*	FromGitter quit (Ping timeout: 260 seconds)
01:10:07	*	oprypin quit (Ping timeout: 260 seconds)
01:43:20	*	apahl quit (Ping timeout: 244 seconds)
01:45:19	*	apahl joined #nim
01:46:39	*	ptdel quit (Quit: https://quassel-irc.org - Chat comfortably. Anywhere.)
01:50:08	*	FromGitter joined #nim
01:53:03	*	endragor joined #nim
02:09:42	FromGitter	<gogolxdong> Is Nim memory safe for now?
02:10:41	*	waleee-cl quit (Quit: Connection closed for inactivity)
02:12:17	FromDiscord	<Varriount> It always has been?
02:12:39	FromDiscord	<Varriount> Or rather, it depends on what you call "memory safe"
02:15:41	FromDiscord	<Elegant Beef> If you turn off the GC it's up to you to make it memory safe 😛
02:17:08	FromDiscord	<Elegant Beef> But that's if we're just talking about leaking of course
02:24:59	*	muffindrake quit (Ping timeout: 246 seconds)
02:27:19	*	muffindrake joined #nim
02:28:33	FromDiscord	<Varriount> @Elegant Beef Or you can use this garbage collection scheme: https://devblogs.microsoft.com/oldnewthing/20180228-00/?p=98125
02:29:58	*	sagax quit (Remote host closed the connection)
02:30:13	FromDiscord	<Elegant Beef> That's beautiful
02:33:14	*	awe00 quit (Ping timeout: 258 seconds)
02:33:16	*	pbb quit (Remote host closed the connection)
02:33:27	FromDiscord	<Varriount> Destroying the hardware: The ultimate garbage collection method.
02:33:39	FromDiscord	<Elegant Beef> And the memory cant be read, so it's safe
02:36:30	*	pbb joined #nim
02:41:16	FromDiscord	<Rosen> one of the only programs that *creates* garbage when it halts perhaps?
02:42:51	FromDiscord	<Elegant Beef> Those missles probably kill civilians and you're saying they're garbage... tsk tsk 😛
02:45:37	*	vicfred joined #nim
03:02:12	*	thomasross quit (Ping timeout: 256 seconds)
03:08:46	*	sagax joined #nim
03:22:32	*	endragor quit (Remote host closed the connection)
03:22:49	*	endragor joined #nim
03:23:35	*	endragor quit (Remote host closed the connection)
03:24:01	*	endragor joined #nim
03:29:01	*	endragor quit (Ping timeout: 264 seconds)
03:32:48	*	endragor joined #nim
03:47:26	*	pbb quit (Remote host closed the connection)
03:50:49	*	pbb joined #nim
03:57:23	*	rockcavera quit (Remote host closed the connection)
04:02:01	*	tsujp quit (Ping timeout: 246 seconds)
04:04:40	*	tsujp joined #nim
04:06:01	*	supakeen quit (Quit: WeeChat 2.9)
04:06:37	*	supakeen joined #nim
04:38:13	*	arecacea1 quit (Remote host closed the connection)
04:38:32	*	arecacea1 joined #nim
04:43:20	*	njoseph quit (Quit: http://quassel-irc.org - Chat comfortably. Anywhere.)
04:43:27	*	njoseph joined #nim
04:58:07	*	hyiltiz quit (Ping timeout: 244 seconds)
05:01:51	*	hyiltiz joined #nim
05:01:51	*	hyiltiz quit (Changing host)
05:01:51	*	hyiltiz joined #nim
05:13:32	*	njoseph quit (Quit: No Ping reply in 180 seconds.)
05:14:37	*	njoseph joined #nim
05:35:32	*	solitudesf joined #nim
05:36:06	*	narimiran joined #nim
05:41:12	silvernode[m]	Good morning. I have been on vacation for a week already and haven't really hung out with everyone here. Why is it, that when I finally get a vacation to actually catch up on hobbies and learning stuff, I procrastinate nearly the entire vacation?
05:41:28	silvernode[m]	ha
05:41:43	silvernode[m]	I suspect I am not alone in that.
05:42:47	narimiran	silvernode[m]: because your vacation is too short
05:43:38	silvernode[m]	<narimiran "silvernode: because your vacatio"> I I think you are correct. I usually get the drive and motivation to do stuff at the end of it.
05:47:42	*	arecacea1 quit (Remote host closed the connection)
05:48:26	*	arecacea1 joined #nim
05:54:23	*	maier joined #nim
05:54:29	*	maier is now known as kenran
05:54:40	*	arecacea1 quit (Remote host closed the connection)
05:55:01	*	arecacea1 joined #nim
06:00:37	*	bung quit (Ping timeout: 258 seconds)
06:06:29	*	NimBot joined #nim
06:12:35	*	oprypin joined #nim
06:29:52	*	bung joined #nim
06:34:25	*	bung quit (Ping timeout: 264 seconds)
06:48:37	*	PMunch joined #nim
06:55:17	Zevv	i had the same. big plans but a total mental cresh, just porridge for brains in the first week
06:55:22	Zevv	eb
06:55:28	Zevv	embrace that
07:13:59	*	Vladar joined #nim
07:14:41	PMunch	silvernode[m], sounds like all of my vacations ever <_<
07:15:25	silvernode[m]	yeah I think we all just want to rest.
07:17:19	PMunch	I wonder how productive I would be if I had all the wealth I would need
07:17:43	PMunch	I.e. if the incentive to work for money was completely gone
07:18:17	*	bung joined #nim
07:32:41	silvernode[m]	<PMunch "I wonder how productive I would "> I have thought deep about this on many occasions and I think after awhile, most of us would absolutely need to be productive. Hungry to get something done and kill the boredem.
07:33:05	PMunch	Exactly
07:33:28	PMunch	But I wouldn't feel the pressure to get up super early in the morning and sit around being an unproductive zombie :P
07:34:12	silvernode[m]	Once you realize that months have gone by and you are free of the shackles, and that you can really work on whatever you want and at your own pace...hobbies would go into overdrive.
07:34:44	silvernode[m]	unless you have a soul sucking spouse. of course.
07:35:07	*	krux02 joined #nim
07:35:11	*	fredrikhr8 joined #nim
07:35:41	PMunch	Well, now you have the energy to get rid of them :P
07:36:37	*	hnOsmium0001 quit (Quit: Connection closed for inactivity)
07:38:45	*	fredrikhr quit (Ping timeout: 258 seconds)
07:38:45	*	fredrikhr8 is now known as fredrikhr
07:45:05	FromDiscord	<Varriount> I would like to both spend time on my hobbies and be paid for it...
07:46:25	PMunch	We should convince some Russian oligarch or something to sponsor development of random Nim projects :P
07:46:45	PMunch	Just give us a couple million and we'll divide us amongst anyone who wants to write Nim
07:48:12	FromDiscord	<Elegant Beef> What is this a work based ponzi scheme
07:48:53	PMunch	Well, not really a ponzi scheme
08:03:40	*	Kaivo quit (Ping timeout: 258 seconds)
08:05:58	*	Kaivo joined #nim
08:18:09	*	Zectbumo quit (Remote host closed the connection)
08:30:35	*	zedeus quit (Ping timeout: 240 seconds)
08:32:54	*	zedeus joined #nim
08:38:28	silvernode[m]	I wan to start a network of audio/video media around Nim and dub them "NimCasts". The network can have a website, and a list of partner Nim programmers that also have audio/media episodes.
08:39:14	silvernode[m]	Anyone who would want to start a show can because part of the network of shows about Nim. This includes live streams too.
08:39:42	FromDiscord	<vieru> is there a way to unallocate a constant ?
08:40:07	PMunch	vieru, nope
08:40:19	PMunch	Constants are a compiler thing, they don't exist on runtime
08:40:33	PMunch	They just get subbed in wherever you use them
08:40:35	FromDiscord	<vieru> sent a code paste, see https://play.nim-lang.org/#ix=2vaO
08:40:39	FromDiscord	<vieru> yeah but they do ocupy memory
08:40:52	PMunch	So `const x = 100; echo x` just gets rewritten to `echo 100`
08:41:27	Oddmonger	yes it's not allocation, it's just pre-computation
08:41:30	PMunch	Yeah, that just gets rewritten to `writeFile("file2", "whatever the contents of 'file' was")`
08:41:55	FromDiscord	<vieru> idk man my file memory usage is 1mb without that code↵and with that code is 7mb
08:42:12	FromDiscord	<vieru> (edit) 'file' => 'executable'
08:42:30	PMunch	Probably because writeFile must allocate that string in memory
08:42:42	PMunch	But it will get collected by the garbage collector as soon as the file writing is done
08:42:43	Oddmonger	the file is included in your exe
08:43:16	PMunch	Is there a built-in way in Nim of echoing variables with their names?
08:43:34	PMunch	I want something like `dump x` to output "x: <value of x>"
08:43:35	krux02	PMunch, not that I know
08:43:45	FromDiscord	<vieru> so you're saying i shoudnt worry about this stuff
08:43:51	krux02	would be very helpful, I agree
08:43:52	FromDiscord	<vieru> gc will do its job ?
08:43:57	PMunch	vieru, exactly
08:43:59	Araq	PMunch, there is macros.dump
08:44:16	FromDiscord	<vieru> thanks
08:44:27	PMunch	I just get undeclared identifier even after importing macros..
08:44:31	Araq	no, there is now strformat
08:45:14	Araq	echo fmt"{expr=}"
08:45:15	*	awe00 joined #nim
08:45:24	PMunch	Eh, that's too terse
08:46:11	Araq	huh?
08:48:27	silvernode[m]	Now that Araq is here, I should ask about official LLVM support and whether or not it is a planned endeavor. I think I saw the ability to use clang somewhere so I guess that counts.
08:50:52	silvernode[m]	I have been looking into LLVM lately and discovered how amazing it is, so naturally I came here to see if my favorite programming language can use it.
08:51:47	FromDiscord	<Clyybber> PMunch: Its sugar.dump
08:52:12	FromDiscord	<Clyybber> silvernode: Theres NLVM
08:52:23	FromDiscord	<Clyybber> which is a LLVM backend for Nim
08:52:36	silvernode[m]	I saw that but it's not official right?
08:53:35	silvernode[m]	<FromDiscord "<Clyybber> which is a LLVM backe"> have you tried it?
08:54:46	FromDiscord	<Clyybber> yeah
08:55:13	FromDiscord	<Clyybber> to both questions :)
08:55:19	silvernode[m]	<FromDiscord "<Clyybber> yeah"> What was you impression of it?
08:55:25	silvernode[m]	oh ok
08:55:33	silvernode[m]	if it's official then I'm in
08:55:40	FromDiscord	<Clyybber> It works :)
08:55:46	FromDiscord	<Clyybber> no, its *not* official
08:55:55	silvernode[m]	aw
08:56:07	silvernode[m]	It would be cool if it was default
08:56:23	FromDiscord	<Clyybber> maybe in the future
08:56:51	FromDiscord	<Clyybber> keep in mind you can still get llvm out of Nim without NLVM
08:57:01	silvernode[m]	The error messages and debugging was awesome with it when I used Rust a few times.
08:57:19	silvernode[m]	<FromDiscord "<Clyybber> keep in mind you can "> is that because of how diverse LLVM is?
08:57:20	FromDiscord	<Clyybber> thats probably unrelated to LLVM
08:57:31	FromDiscord	<Clyybber> its because you can compile C to LLVM
08:58:24	silvernode[m]	yeah I was old by someone (who probably didn' know wha they were talking about) that the error messages in rust were from LLVM. I guess that's BS
08:59:19	silvernode[m]	If we can get that level of detail with our error messages in Nim, it would be fantastic.
08:59:50	silvernode[m]	That's one of the biggest hurdles I have faced while learning Nim.
09:00:49	FromDiscord	<Elegant Beef> Yea i just ran into a pretty annoying problem i feel an error message would've been nice, trying to use methods without the base type imported should tell you, import the god damned base type 😄
09:02:14	FromDiscord	<Elegant Beef> I spent vastly more time than i'd like to admit trying to figure out why that happen, eventually i just tried include and realize that the only difference with the include vs import is that a single type wasnt present
09:02:48	PMunch	Aah, thanks @Clyybber
09:05:55	silvernode[m]	<FromDiscord "<Elegant Beef> I spent vastly mo"> Yeah disruptek asked me back in Feb when I first started using Nim what I think is the biggest problem for adoption of Nim. I said whitespace, but after programming with Nim for the better part of a year now, I think the error messages are a actual issue.
09:06:19	silvernode[m]	Whitespace it not a problem
09:06:52	Araq	error messages are generally good these days, what in particular did you encounter?
09:08:07	FromDiscord	<haxscramper> Error messages are not that good when it comes to errors in template-generated code. It takes just couple of `mapIt().filterIt()` to generate several screens worth of text. Type mismatches on function types are also could be improved
09:08:20	silvernode[m]	I would trade anything to have detailed error messages. I otherwise we spend hours going crazy just like you did
09:08:37	FromDiscord	<haxscramper> I checked how error messages in compiler are passed to the user - right now everything is basically squeezed into single string and passed over to message printer. Well, except for type mismatch errors where it first collects all alternatives and then prints them.
09:08:50	Zevv	hexscramper: can you come up with concrete examples of bad error messages?
09:08:52	FromDiscord	<Elegant Beef> If that is to me I had sub types declared in seperate files and the master type in a file i didnt import in the main module where the method call was happening so the conversion from Controller -> AiController or Controller -> PlayerController needed to be explicit
09:09:05	silvernode[m]	<Araq "error messages are generally goo"> Usually it's "instantiation of fmt from here" messages
09:09:32	FromDiscord	<Elegant Beef> And ther was no indication that the base class needed to be imported for the methods to work
09:09:40	silvernode[m]	I think that messages in particular needs work
09:09:58	silvernode[m]	* I think that message in particular needs work
09:10:37	silvernode[m]	Araq: Overall though, the error messages are pretty good.
09:10:42	Araq	<Elegant Beef> I think that's a language wart moreso than a problem of error messages
09:11:04	Araq	I keep asking for attaching procs to types, not many are too happy about it
09:11:08	FromDiscord	<haxscramper> Zevv: https://play.nim-lang.org/#ix=2vb3
09:11:22	FromDiscord	<haxscramper> It approaches C++ level of template-error-maddness
09:11:32	FromDiscord	<Elegant Beef> Well either way it was unclear the problem and how to resolve it, what do you mean attaching procs to type, like member functions?
09:11:40	FromDiscord	<haxscramper> Yes, not everyone writes tonts of templates like that
09:12:05	silvernode[m]	<FromDiscord "<haxscramper> Zevv: ">https://play"> I didn't know you could put discard there 🙂
09:12:12	FromDiscord	<haxscramper> But showing /original/ source code is much better than generated-one
09:12:33	FromDiscord	<Elegant Beef> I was actually originally using a proc on the base type and converting to the required type inside the proc, which works equally well as methods i feel, just since methods were there though they were "more proper" for this type of inheritance OOP
09:12:50	FromDiscord	<Elegant Beef> thought*
09:14:15	Zevv	hexscramper: there is a lot of noise in there, but the first line is spot on, right?
09:14:22	*	Trustable joined #nim
09:15:13	Zevv	but you're right, having expanded code in the error messages is not helpful here
09:15:46	FromDiscord	<haxscramper> Yes. Like, /technically/ this is not an issue whatsoever. Just there is too much noise
09:15:49	Zevv	it approaches C++ level of template-error-madness
09:16:08	FromDiscord	<haxscramper> Actually, I looked how messages are dealt with in compiler
09:16:26	Zevv	they're pretty flat
09:16:39	FromDiscord	<haxscramper> And maybe I will try to do something with it? I'm not sure if I can pull that off, but I want to at least try
09:17:02	Zevv	but what is the root problem then
09:17:49	FromDiscord	<haxscramper> Ideally I want to have at least something like that - https://github.com/haxscramper/hmisc#hmischexceptions-documentation - pretty hacky right now, but works somewhat
09:19:44	Zevv	its cool for some, but not for others. I can't stand the "modern" gcc errors with soo much arrows and white space and colors
09:20:19	FromDiscord	<Elegant Beef> Yea just tell me why im a dullard and get on with it
09:21:04	FromDiscord	<Elegant Beef> Like do we really need arrows to say "You fucked up here"
09:21:17	FromDiscord	<haxscramper> If there is concrete type for reporting errors in code (~CodeError~) it is just a matter of representation - pretty-pring everying in the terminal, output old-style errors in cases where it is necessary, generate json ouput for automatic processing etc.
09:22:27	FromDiscord	<haxscramper> Right now it all happens using strings which means there is no configuration possible at all.
09:22:37	FromDiscord	<haxscramper> I'
09:23:34	FromDiscord	<haxscramper> I'm not really sure if I understood the logic of error handling in compiler, but from what I saw - various procedures just call 'handle error' proc that prints things, and optionally quits.
09:24:31	FromDiscord	<Clyybber> also are we talking compile error messages or runtime?
09:24:37	FromDiscord	<Clyybber> silvernode ^
09:25:43	FromDiscord	<haxscramper> Why not raise exception for fatal compilation errors with absolutely all information necessary. This is personal preference of course, but when error happens I usually don't even think about how expensive it would be to print nice message - there is nothing else to do anyway.
09:27:27	FromDiscord	<Rika> i like the error arrows, otherwise i wouldnt notice that i was looking at the wrong line
09:27:33	FromDiscord	<Rika> or the wrong portion or the line
09:27:44	FromDiscord	<Rika> colors on errors are eh, they look cool but thats it
09:28:46	FromDiscord	<Elegant Beef> I'd rather just capture the snippet of the error and display it in one line with line number/file on the side
09:29:00	Zevv	we can ask disruptek to do it, sure he'll find a matching rainbow colored animated dingbat for each possible error
09:29:16	FromDiscord	<Elegant Beef> Well it might make his balls hurt though
09:29:20	FromDiscord	<Elegant Beef> Cant think with throbbing balls
09:30:05	*	zedeus quit (Ping timeout: 240 seconds)
09:30:12	FromDiscord	<haxscramper> > I'd rather just capture the snippet of the error and display it in one line with line number/file on the side
09:30:25	FromDiscord	<Elegant Beef> No too much noise imo
09:31:10	FromDiscord	<haxscramper> If I remove errors it would be better? Just leave upper part of the message + source code
09:31:46	FromDiscord	<Rika> that's too vertically long for me
09:31:53	FromDiscord	<Elegant Beef> `file.nim line# (wrong code)`↵`optional arrow`↵`Error description`
09:32:00	FromDiscord	<Elegant Beef> That's somethin i can get behind
09:32:08	*	zedeus joined #nim
09:32:14	FromDiscord	<Elegant Beef> at minimum 2-3 lines of error
09:33:20	FromDiscord	<Rika> sent a code paste, see https://play.nim-lang.org/#ix=2vb7
09:33:38	FromDiscord	<haxscramper> AFAIK rust has all possible compiler errors indexed - compiler just shows simple error and there is a webiste with long description. Is it a good idea?
09:33:40	FromDiscord	<Elegant Beef> I dont like the multiple erros in one area
09:33:50	FromDiscord	<Rika> its not, its part of the same error
09:33:53	FromDiscord	<Rika> just more context
09:34:28	FromDiscord	<Elegant Beef> So... it's noise
09:34:33	FromDiscord	<Rika> > AFAIK rust has all possible compiler errors indexed - compiler just shows simple error and there is a webiste with long description. Is it a good idea?↵@haxscramper no, unless they also link to the (local installation of the) site in the error
09:34:46	FromDiscord	<Rika> i sometimes program in no-internet situations
09:35:02	FromDiscord	<Elegant Beef> `$a` isnt the issue `@a` is the way i see it in their example
09:35:03	FromDiscord	<Rika> > So... it's noise↵@Elegant Beef everything is noise to you it feels
09:35:14	FromDiscord	<haxscramper> But if there was such installation it is acceptable?
09:35:37	FromDiscord	<Rika> i think so, as long as the link is there too
09:36:11	FromDiscord	<Elegant Beef> it sincerely took me like 5 seconds to extract information from that screenshot, although it's 3:30am here
09:36:41	FromDiscord	<haxscramper> > `$a` isnt the issue `@a` is the way i see it in their example↵@Elegant Beef Need more colors!!! :)
09:36:46	FromDiscord	<Rika> i think the indentation is important, makes it easier for us humans to determine what lines are related to each other
09:36:52	FromDiscord	<Elegant Beef> I went to the arrows first, and went nope, then went to to the top and nope, then went to the error finally
09:37:20	FromDiscord	<haxscramper> Oh, yes, arrows like x3 times more flashy then the error itself.
09:37:23	FromDiscord	<Elegant Beef> It needs the actual error looking like an error being red and the context arrows being white unless it's error
09:37:46	FromDiscord	<Elegant Beef> Multiple arrows made the error ambigious to where it was
09:37:57	FromDiscord	<Elegant Beef> `$a` isnt the error, `@a` is
09:39:57	FromDiscord	<Elegant Beef> Anywho i sleep forward all my hatemail to your least favourite person
09:40:18	FromDiscord	<haxscramper> To sum up: actual error message should be red, arrow annotations white/dim. Don't put too many arrows all over the place
09:40:19	FromDiscord	<Rika> what if that person is you
09:40:43	FromDiscord	<Elegant Beef> Rika at this point i'd just be happy i made an impression on someone
09:40:45	FromDiscord	<Rika> hax: not always possible when using older terminal
09:40:45	FromDiscord	<Elegant Beef> 😄
09:41:06	FromDiscord	<Rika> dont recognize beef as a human, got it
09:42:25	FromDiscord	<haxscramper> There should be like 600+ page book on 'error printing best practices' or something and seems like I need to read it
09:43:33	FromDiscord	<Rika> if we're gonna go all in with color why even bother with arrows; color the bg red for errors or something
09:43:44	FromDiscord	<--HA--> sent a long message, see http://ix.io/2vb9
09:44:33	FromDiscord	<Rika> isnt `void*` == `pointer`
09:45:25	FromDiscord	<Rika> i dont see how the `type pointer with calling it with nimstrvar[0].addr` would fail
09:45:35	FromDiscord	<--HA--> That is what I found in a forum post, too. I tried this basically <https://forum.nim-lang.org/t/3045>
09:46:47	FromDiscord	<--HA--> sent a code paste, see https://play.nim-lang.org/#ix=2vba
09:48:36	FromDiscord	<Rika> perhaps? also i think you dont need to convert to cstring
09:58:55	FromDiscord	<--HA--> That also happens implicitly, yes. I tried some variations but never got a positive result.
10:04:50	FromDiscord	<Rika> maybe check if magick is "returning an exception" via its exceptions api
10:06:51	FromDiscord	<--HA--> That is a good idea, thanks @Rika
10:14:11	*	zielmicha__ quit (Read error: Connection reset by peer)
10:14:43	*	rayman22201 quit (Read error: Connection reset by peer)
10:14:43	*	alehander92 quit (Read error: Connection reset by peer)
10:15:20	*	l1x quit (Read error: Connection reset by peer)
10:15:33	*	rayman22201 joined #nim
10:15:34	*	zielmicha__ joined #nim
10:15:39	FromDiscord	<krisppurg> Is Nim beginner friendly for new programmers?
10:15:46	*	l1x joined #nim
10:15:46	*	alehander92 joined #nim
10:16:56	FromDiscord	<ache of head> Hmm
10:17:04	FromDiscord	<ache of head> I'd say there are languages more suited for beginners
10:17:38	FromDiscord	<ache of head> But, for example, I'd say I'd learn Nim as a beginner more easily than, say, Java
10:18:46	FromDiscord	<Rika> Not as much as python and javascript is
10:18:55	FromDiscord	<Rika> Documentation is everything for beginners
10:19:12	FromDiscord	<Rika> And nim is fairly lacking there
10:19:29	FromDiscord	<Rika> (I mean Documentation in the sense of tutorials guides and stuff as well(
10:20:05	*	awe00_ joined #nim
10:20:09	silvernode[m]	<FromDiscord "<krisppurg> Is Nim beginner frie"> I think it could be good for beginners since they don't have expectations but on the other hand, there are easier languages they could start with. In the end it depends on the person and their intention. Everyone's brain works differently.
10:20:15	FromDiscord	<haxscramper> I think nim is a good language for beginners in programming actually. Documentation is lacking when it comes to relatively high-level stuff. Things like types/iterations etc. are described good enough in the manual
10:20:18	PMunch	Ugh, I think I need some sleep... Really can't do math today
10:21:15	PMunch	I'd say Nim is very beginner friendly because of the helpful community
10:21:23	silvernode[m]	<PMunch "Ugh, I think I need some sleep.."> 1 sheep x 10 sheep x 50 sheep x 100 sheep.
10:21:25	*	awe00 quit (Ping timeout: 240 seconds)
10:21:32	PMunch	If you're stuck on something you can almost always just ask and get a good reply
10:21:55	FromDiscord	<haxscramper> And there is like, close to zero of things like `__name__ == 'main'`, special variables end things when you need to keep in mind when starting in python. At least for me it is somewhat confusing
10:22:02	FromDiscord	<Rika> Beware of dis___tek though ;)
10:22:09	PMunch	Haha :P
10:22:30	silvernode[m]	<PMunch "If you're stuck on something you"> yeah I think that is the best point yet. I have seen beginners give up because they go into the python irc and there are just so much going on that they get ignored.
10:22:33	FromDiscord	<Rika> Ah yeah I always have to search the proper magic variable or method in python
10:22:51	FromDiscord	<Clyybber> you ask disruptek how to echo a variable and you get back a reflection on the nature of the universe, data and mutations
10:23:51	FromDiscord	<Clyybber> what more is disruptek than an abstract entity thinking about abstract entities with throbbing balls?
10:24:55	FromDiscord	<haxscramper> To check if `distinct ptr[T]` is non-nil I first need to cast it back to `ptr`, right?
10:25:14	silvernode[m]	Maybe beginners should start with assembly and move to Nim once they can't take the pain. That way they appreciate it and stick with it when they have problems. 🙂
10:25:59	FromDiscord	<Recruit_main707> haxcramper, maybe you can borrow the isNil function
10:26:24	PMunch	Anyone wants to help me write my simple iterator? Basically I have an array that I want to use as a circular buffer. By passing in two exclusively increasing indices (stop always larger than start) I want to essentially get all the values for that range mod the sequence. I had it working just fine, but I'm struggling with if you give a sequence like 0..15 on a buffer of size 10, that should repeat the first five elements again..
10:26:49	FromDiscord	<Recruit_main707> or create a template `==` that casts itfor you at least
10:27:28	FromDiscord	<Rika> disruptek is great but newbies must not interact with them or they may face severe consequences
10:29:26	silvernode[m]	<PMunch "Anyone wants to help me write my"> "circular buffer", I don't think I'd be of much use there. That must be deeper into programming than I have gone yet.
10:29:55	PMunch	God damn it...
10:30:09	PMunch	I've spent hours writing this thing, and just realised how to do it in three lines..
10:30:18	PMunch	Nvm, two lines..
10:30:19	FromDiscord	<Rika> lol
10:30:33	FromDiscord	<Rika> using a modulo
10:30:48	FromDiscord	<Rika> and a for loop over a range
10:30:50	silvernode[m]	<PMunch "I've spent hours writing this th"> That's good though, since you get a valuable lesson.
10:30:55	PMunch	for i in start..stop: field buffer[i mod array.len]
10:31:29	silvernode[m]	I love it when that kind of stuff happens to me. It makes me a better programmer.
10:31:33	PMunch	@Rika, exactly...
10:31:46	FromDiscord	<Rika> rip lol
10:31:57	PMunch	I had all this complex logic about splitting the range into subranges and such
10:32:00	PMunch	A complete mess
10:32:17	silvernode[m]	I always do that to myself
10:32:29	silvernode[m]	that's why I come here and ask dumb questions all the time.
10:32:51	PMunch	Even struggled so much that I wrote an auto-test procedure to test a bunch of different combinations
10:33:35	silvernode[m]	<PMunch "Even struggled so much that I wr"> and this is why my text game is going to be a complete mess.
10:33:48	PMunch	Haha :P
10:34:07	silvernode[m]	over thinking is one of my biggest problems as a programmer.
10:34:45	silvernode[m]	when you spend days on something only to have someone look at your code and point out the solution in 5 seconds.
10:42:07	*	abm joined #nim
10:47:19	*	stisa[m] joined #nim
10:58:00	FromGitter	<alehander92> haxscramper
10:58:03	FromGitter	<alehander92> i love this
10:58:19	FromGitter	<alehander92> i also want to work on elm-like errors in some of my own projects
10:58:27	FromGitter	<alehander92> so i might follow this
10:58:31	FromDiscord	<haxscramper> Error message i pasted?
10:58:48	FromGitter	<alehander92> the idea of having more "fancy" messages*
11:00:08	FromDiscord	<haxscramper> Right now I'm just using quickly hacked together library, but I will probably move it into separate one + add suggestions from discussion above - configurable printing of error messages.
11:10:52	*	theelous3 joined #nim
11:12:14	FromDiscord	<--HA--> @Rika The image magick error is NoDecodeDelegateForThisImageFormat and it gave me a line in the c code. If I pass in an empty string I get the ZeroLengthBlobNotPermitted error I could see existed from the c source. So something is being passed to the function but no matter what I put in there the image format is not recognized. Read from a file directly works. Reading that same file into a string and passing it to MagickReadImageBlob doesn
11:14:15	FromDiscord	<Rika> prolly the magic header is unrecognized?
11:14:31	FromDiscord	<Rika> let me read the docs one moment
11:16:13	FromDiscord	<--HA--> It looks more like it is empty. I'm not very familiar with c but it seems like the magic header should be included in the error message and there is just nothing. And using the function MagickReadImage recognizes it fine. My best guess atm is that somehow I pass the data in a wrong way.
11:16:27	FromDiscord	<Rika> maybe
11:16:52	FromDiscord	<--HA--> I got MagickGetImageBlob working and passing that blob back to MagickReadImageBlob also doesn't work
11:17:33	FromDiscord	<Rika> i dont know what the issue would be, sorry
11:18:21	FromDiscord	<--HA--> Thank you for looking into it!
11:19:44	shashlick	Which wrapper are you using
11:20:21	shashlick	Also snippet may help
11:21:57	FromDiscord	<--HA--> It's posted above a bit. The discord link to the message probably does not help for IRC. No wrapper, if I understand the question correctly. I wrote a few proc with importc pragma.
11:26:46	*	lritter joined #nim
11:35:03	FromDiscord	<--HA--> I've been looking a lot at <https://github.com/zedeus/nimagemagick> but there is no read blob function there yet.
11:45:01	*	rockcavera joined #nim
11:45:37	PMunch	Ugh, another math thing that's taking me waay to long to figure out
11:46:24	PMunch	I'm creating a particle system
11:46:55	PMunch	Particles should be able to move fast or slow, their x and y speed is stored as two int8s
11:47:11	PMunch	I want the slowest speed to be sub 1 pixel per frame
11:47:27	FromDiscord	<lqdev> make the velocity fixed point?
11:47:37	PMunch	Let's say minimum speed is one pixel every four frames or something
11:47:37	FromDiscord	<lqdev> and position ofc
11:47:48	PMunch	Hmm, how would that work?
11:48:02	FromDiscord	<lqdev> then you shift whole numbers two bits to the right
11:48:24	FromDiscord	<lqdev> the two least-significant bits are then your representations of .0, .25, .5, and .75
11:48:30	PMunch	I was thinking just calculate the "I need to move x pixels every n frames" and then do a "if frame mod n: pos + x"
11:48:35	FromDiscord	<Rika> @--HA-- well because the wrapper is generated automatically, the code isnt there
11:48:41	FromDiscord	<lqdev> PMunch: mod is slow
11:48:41	FromDiscord	<Rika> but it will be when you use the library
11:48:50	FromDiscord	<lqdev> if you're using ints for speeds you're likely constrained
11:48:57	FromDiscord	<lqdev> so do not use mod
11:49:01	FromDiscord	<lqdev> unless you're not constrained
11:49:07	FromDiscord	<lqdev> then i don't see a reason for not using floats
11:49:46	PMunch	Well I'm mostly constrained in program space
11:50:00	PMunch	As in the size of the program and the amount of stuff I can keep in RAM
11:50:08	FromDiscord	<lqdev> what about clock speeds?
11:50:34	FromDiscord	<lqdev> because `mod` is like, suuuuuuuper slow, esp. if your processor doesn't have an instruction for it
11:50:40	FromDiscord	<lqdev> same with division
11:50:54	shashlick	@--HA-- cstring is already a pointer
11:51:53	PMunch	It's a 16Mhz chip
11:51:56	shashlick	You could just pass blob.cstring
11:51:59	FromDiscord	<lqdev> hm
11:52:04	PMunch	But I'm not 100% sure if I'm running it at that speed atm
11:52:10	FromDiscord	<lqdev> and is this in a tight loop?
11:52:28	PMunch	Ah, yes it's running at 16Mhz
11:52:40	PMunch	Yeah, but I'm already using mod and div all over the place
11:52:44	FromDiscord	<lqdev> oof
11:52:46	PMunch	Speed doesn't seem to be much of an issue
11:53:07	PMunch	It's just the images and level data that's filling up the puny amount of memory on this thing
11:53:42	PMunch	28672 bytes of program memory, and 2560 bytes of data memory
11:56:06	FromDiscord	<lqdev> fixed-point numbers are really simple: instead of representing your values as whole units, you represent it as 1/4 units (you assign two bits to the fractional part, so the format is 6.2 fixed-point, as in integer_bits.fractional_bits)
11:56:15	FromDiscord	<lqdev> well 1/4th
11:56:18	FromDiscord	<lqdev> it can be any fraction
11:56:24	FromDiscord	<lqdev> as long as the denominator is a power of two
12:05:37	FromDiscord	<--HA--> I just got it working with the inbuilt "logo:" image and using a different format. Thanks everyone for your help! Turns out the issues seems to be just with detection of the image type I'm working with, which is TGA. But I know it generally works now and my wrapper code is not the issue.
12:06:02	*	supakeen quit (Quit: WeeChat 2.9)
12:06:22	*	pbb quit (Ping timeout: 260 seconds)
12:06:24	FromDiscord	<Rika> nice
12:06:37	*	supakeen joined #nim
12:08:53	*	waleee-cl joined #nim
12:12:06	PMunch	lqdev, this is very promising!
12:14:01	*	pbb joined #nim
12:14:56	Zevv	make sure to shift back after multiplying :)
12:16:33	PMunch	Shift back?
12:16:43	FromDiscord	<lqdev> when rendering i guess
12:16:56	PMunch	I need to use everything <1 as every n frames add another pixel
12:16:59	FromDiscord	<lqdev> you have to be careful 'cause now 1 means 1/4th of a pixel and not a single pixel
12:17:45	PMunch	I guess I could create a procedure that takes a static[float] and converts it automatically
12:18:39	FromDiscord	<lqdev> distinct types to the rescue!
12:29:45	*	Vladar quit (Remote host closed the connection)
12:36:11	Zevv	pmunch: you can just add or sub your 6.2s, but if you multiply then, you get a 12.4, which you have to shr by 2 to make it a .2 again
12:40:01	PMunch	Uhm..
12:40:49	Zevv	for example, 3.25*5.75 would be 13*23=299. you do 299 shr 2 to get your answer = 75, meaning 18.75 in 6.2
12:40:52	Zevv	does that make sense?
12:41:10	Zevv	but maybe you dont ever need to multiply, so then it doesnt matter :)
12:41:35	PMunch	Ah right, I see
12:41:43	PMunch	No I don't think I need to multiply :P
12:41:51	Zevv	until you do!
12:41:53	PMunch	But having a fixed-point library for Nim would be cool
12:42:18	Zevv	there must be a few
12:43:50	*	FromDiscord quit (Remote host closed the connection)
12:44:06	*	FromDiscord joined #nim
12:44:09	Zevv	well, there seem to be none. pmunch to the rescue!
12:44:48	PMunch	I'm barely awake enough to make myself coffee, let alone writing complex bit-shift mathy libraries!
12:45:41	FromDiscord	<Rika> speaking of weird decimal math, i think i made a float16 library some time ago
12:46:53	PMunch	Hmmmm
12:47:10	PMunch	I guess I could also store the positions as fixed-point..
12:47:18	PMunch	That would solve a lot of my issues..
12:47:40	Zevv	well, if you want your positions to increment sub-int, I guess you should right
12:47:47	PMunch	Yeah
12:48:28	PMunch	My plan was to use a scheme of adding another pixel every N frames to simulate it
12:48:36	Zevv	but with 6.2 you can only go up to 63.75
12:48:38	PMunch	And keeping the position an integer
12:48:49	Zevv	nah way too complex
12:49:11	PMunch	I store the position as an int16 at the moment
12:49:28	PMunch	So that would be a 14.2
12:49:29	Zevv	so that would be a 14.2
12:49:40	Zevv	dude will you marry my. disruptek refused
12:49:43	PMunch	Haha, exactly
12:50:29	PMunch	Oof, long distance marriage?
12:50:52	Zevv	who cares, it's 2020 and we have irc
12:50:57	PMunch	Haha :P
12:51:16	Zevv	let me see how nice generic fixedpoint can be put in some type system
12:52:09	PMunch	I mean this project is part of a gamejam to make a pretty and simple game. With extra points for simple source code
12:52:29	PMunch	I'm going hard for the simple source code thing with some crazy Nim macros
12:52:44	Zevv	hehe
12:52:55	PMunch	Showing them generic fixed point numbers would definitely be cool
12:56:02	PMunch	Hmm, what's the best way to do rounding?
13:02:54	Zevv	to whole numers?
13:02:58	PMunch	Yeah
13:03:02	Zevv	just like normal floats. add 0.5 and truncate
13:03:08	PMunch	Aaah
13:03:10	PMunch	Ofc
13:03:12	Zevv	assuming > 0 of c
13:03:47	PMunch	Are you making a library now by the way?
13:05:26	Zevv	fiddling a bit
13:05:34	Zevv	to see if I can make something generic
13:05:38	PMunch	Nice
13:05:44	Zevv	so you can say, my T is a 6.2
13:05:49	Zevv	change that to 5.3
13:05:52	Zevv	and everything just works
13:06:59	Zevv	I made a few in C over the years for customers
13:07:03	Zevv	but I can't find any in my own repos
13:07:32	PMunch	That's amazing!
13:07:49	PMunch	Wait, not on runtime?
13:08:04	Zevv	?
13:08:18	PMunch	You can't change a 6.2 to a 5.3 on runtime
13:08:38	Zevv	well, you could convert between them. but I'm trhing to make types for those
13:08:45	PMunch	Yeah okay
13:08:51	Zevv	so you define your speed as a FixedPoint[uint8, 2]
13:08:56	PMunch	Yeah
13:08:59	Zevv	and your location as a FixedPoint[16, 2]
13:09:12	PMunch	And then just location += speed
13:09:27	Zevv	right. But that's already where the complexity lies - mixing them
13:09:40	Zevv	what if I add a 16.4 to a 8.2. what should the output be?
13:09:43	PMunch	Add a round function and a compile-time float -> fixed converter and that's pretty much everything I need :P
13:10:05	PMunch	I'd say that's not possible
13:10:12	PMunch	Without a converter
13:10:17	Zevv	well, `+=` should work of course
13:10:21	FromDiscord	<mratsim> I have some fixed point arithmetic hidden
13:10:25	PMunch	Yeah, but only with the same base
13:10:36	PMunch	Well, same fix point
13:10:45	FromDiscord	<mratsim> https://github.com/mratsim/trace-of-radiance/blob/master/trace_of_radiance/io/color_conversions.nim#L105-L106
13:10:59	PMunch	So taking a 14.2 and adding a 6.2 should work
13:11:42	Zevv	I guess so
13:12:10	Zevv	Oh I thought mratsim already came spoiling our fun with a complete fp math lib
13:12:11	Zevv	but no :)
13:12:16	*	Zevv waves to mratsim
13:12:30	PMunch	Well I guess that could also require conversion. But it would be the same as just doing int16(7'u8)
13:12:58	FromDiscord	<mratsim> Nah, I don't have enough time for that
13:13:11	Zevv	ain't nobody got time for that!
13:13:20	FromDiscord	<mratsim> I want to wrap up my crypto lib so that it starts being useful and then I want to explore multithreaded async
13:13:36	Zevv	ain't nobody got time for that either!
13:14:17	FromDiscord	<mratsim> Luckily there is a very fresh article on Rust futures which I find quite elegant and actually "zero cost abstraction": https://axelforsman.tk/2020/08/24/rust-style-futures-in-c.html
13:14:26	FromDiscord	<mratsim> the article is on how to implement them in C
13:15:02	Zevv	Aw don't let disruptek hear, he's all set to make CPS the next big thing
13:15:09	FromDiscord	<mratsim> Also those futures doesn't involve the GC which makes them threading friendly
13:15:40	Zevv	ooh that looks interesting indeed
13:15:51	FromDiscord	<mratsim> I don't mind exploring CPS either but before that I need a good handle on the high level problem
13:15:56	FromDiscord	<mratsim> threading done.
13:16:01	Zevv	\o/
13:16:03	FromDiscord	<mratsim> Next is async and iterators
13:16:29	Zevv	also cps!
13:18:00	FromDiscord	<mratsim> I mean, the 3 use cases of CPS is unifying threading, async, iterators and in general everything that uses callbacks
13:18:25	Zevv	coroutines
13:18:32	Zevv	but, yes
13:18:46	FromDiscord	<mratsim> once I have an idea of how to write those libraries, I would have a better idea on the magical solution 😉
13:19:10	FromDiscord	<mratsim> if it is the magical solution*
13:19:28	FromDiscord	<mratsim> ah yes macro composition as well. But I think I'm OK on the macro front
13:19:31	Zevv	disruptek got CPS to a kind-of-working state. We can do async like stuff already, implement iterators and coroutines, implement longjmp style gotos, exception handling, etc
13:19:35	FromDiscord	<Admin> How would i change the available elliptic curves for a http client ?
13:21:19	FromDiscord	<stisa> Pmunch I just saw this for fixed point numbers https://gitlab.com/lbartoletti/fpn not sure how complete it is tho
13:22:10	PMunch	Hmm, not the prettiest solution
13:22:11	Zevv	it's not very generic, comes with a few predefined types only
13:22:16	Zevv	and it has a sucky api
13:22:24	PMunch	Yeah, was about to say
13:22:48	PMunch	Seems almost like a port of some C library
13:26:11	FromDiscord	<Admin> > How would i change the available elliptic curves for a http client ?↵anyone have a clue ?
13:26:17	Zevv	bwah, my converter does not like to be [T, W] generic
13:26:59	Zevv	don't know Admin, if it's possible it should probably in the 'net' docs
13:27:14	Zevv	not sure how much of openssl is actually exported into Nim land at this time
13:29:09	FromDiscord	<mratsim> genotrance / shashlick has a full export in nimssl
13:29:39	Zevv	is that mixable with the normal (async) net?
13:30:30	FromDiscord	<mratsim> I think it's all setup in the sslcontext and one of the init function accepts that context
13:30:38	Zevv	ah right
13:30:39	FromDiscord	<mratsim> but you might need a cast
13:31:04	FromDiscord	<mratsim> because the highlevel type name probably differ
13:32:45	shashlick	i don't think nimssl is functional yet
13:33:45	shashlick	i just point those who ask to my rsa.nim test in nimterop or the pkcs7 wrapper here - https://gist.github.com/genotrance/53d3ae50434ad73365d45bdd6a0ff97d
13:36:32	shashlick	i felt openssl was too large to wrap everything but now that you can generate an output file, maybe it's worth it
13:37:21	Zevv	openssl is hell. all of it.
13:37:52	Zevv	mratsim should just put the crown on his crypto lib by making us a nice nim native TLS1.[23] implementation
13:38:34	FromDiscord	<mratsim> Note that we are undergoing security audit and are using BearSSL for basic crypto: https://github.com/status-im/nim-bearssl↵- https://github.com/status-im/nim-libp2p/tree/master/libp2p/crypto
13:39:48	Zevv	excellent choice
13:40:01	FromDiscord	<mratsim> I don't know if I want to implement TLS. It's a super complex beast, it's been done to death, it's unlikely to attract people to Nim because of BoringSSL, BearSSL, WolfSSL, mbedTLS, ...
13:40:35	Zevv	true, I was not being serious. But I *do* feel that it might pay to take a secure and compact SSL lib and make it part of the nim stdlib
13:40:46	FromDiscord	<mratsim> i.e. underexplored crypto with no big libs that is complex to implement and complex to optimize is a better way to attract people to Nim
13:40:53	FromDiscord	<mratsim> ah yes sure.
13:41:15	FromDiscord	<mratsim> But I don't see how that's possible.
13:41:33	Zevv	I also think that good crypto primitives with a a solid API should be part of the core stdlib, but so many people with so many opinions on that
13:42:11	Zevv	well, as long as we target C/C++ we could just move any lib into the stdlib as long as the license agrees?
13:42:28	FromDiscord	<mratsim> The issue with crypto is that we want it to be both fast (HTTP server bottlenecked by crypto) and secure.
13:43:07	FromDiscord	<mratsim> A pure Nim one would need assembly for both to ensure the compiler doesn't introduce branches that would leak secrets
13:43:15	FromDiscord	<mratsim> Assembly like this: https://github.com/mratsim/constantine/blob/master/constantine/arithmetic/limbs_asm_montmul_x86.nim#L127-L188
13:43:21	Zevv	no but pure nim is not soemthing we would want I think
13:43:30	FromDiscord	<mratsim> and this is totally contrary to the stdlib philosophy so far
13:43:37	Zevv	let the security people do their security thing in their libs
13:43:46	FromDiscord	<mratsim> and I don't see how many people would be willing to maintain this
13:43:49	Zevv	we could pull bearssl into the stdlib, technically?
13:43:54	FromDiscord	<mratsim> yes
13:43:59	Zevv	I would so vote or that
13:44:01	Zevv	/for/
13:44:41	FromDiscord	<mratsim> but a crypto API in the stdlib is actually simple
13:45:05	Zevv	sure, but if we have something like bearssl, any crypo primitives could lie on top of that as well
13:45:13	Zevv	no need to have nim implementations?
13:45:23	FromDiscord	<mratsim> you need "newContext" "generateSecretKey", "derivePublickey", "sign", "verify"
13:45:43	FromDiscord	<mratsim> and you can use OpenSSL or BearSSL behind
13:45:43	Zevv	and some things for symmetrical cryoto of course and hashes
13:45:46	Zevv	right
13:46:21	Zevv	but still nothing like that exists, it seems
13:46:38	Zevv	there's 23 differetn nim crypto libs that all do just one or two things, and all have different APIs and most are not maintained
13:46:46	Zevv	I don't dare touch any of them and just bind EVP
13:47:20	FromDiscord	<mratsim> This is what we are doing there:↵- https://github.com/status-im/nim-blscurve/blob/master/blscurve/blst/bls_sig_min_pubkey_size_pop.nim↵- https://github.com/status-im/nim-blscurve/blob/master/blscurve/miracl/bls_signature_scheme.nim
13:47:39	FromDiscord	<mratsim> 2 backends, same API
13:47:59	Zevv	right. all that stuff would look so nice in nims stdlib
13:48:02	shashlick	so is there value in a full openssl wrapper
13:48:44	FromDiscord	<mratsim> only if it can be used in Async net I think
13:49:02	PMunch	Argh, why can't I use a procedure that takes a static[float] inside a macro..
13:49:04	FromDiscord	<mratsim> if you also need to rewrite async net, I don't think there is that much value, too much friction anyway
13:49:27	shashlick	well, I don't know what is missing in async net ssl wise so no idea
13:49:40	FromDiscord	<mratsim> @PMunch, remove static and tag {.compileTime.}
13:49:56	shashlick	i'm just not sure we need a 10mb nim wrapper that covers all of ssl
13:49:58	FromDiscord	<mratsim> it's like const in macros
13:50:21	PMunch	Hmm, that works
13:52:07	*	WilhelmVonWeiner joined #nim
13:52:20	FromDiscord	<mratsim> in a VM context, there is no const or static, just let and plain values.
13:53:02	FromDiscord	<mratsim> This is annoying when you want to do `static: debugEcho ...` in a proc that is called at both runtime and compile-time
13:53:03	PMunch	Hmm, negative values doesn't seem to work very well with my fixed point implementation..
13:53:52	FromDiscord	<mratsim> Floats are sign-and-magnitude based and ints are two-complement
13:54:08	WilhelmVonWeiner	should one use `proc main = ` over writing their code at the top level of a file?
13:56:18	FromDiscord	<ache of head> AFAIK Nim uses `when isMainModule:`
13:57:54	PMunch	WilhelmVonWeiner, it gives a very slight performance boost, but it won't be called implicitly
13:58:25	PMunch	`when isMainModule` is used to run code only when that file is the main module being compiled
13:58:34	PMunch	So you can have a tool and a library in the same file
13:58:37	shashlick	i might enable nimterop to split wrapper output into multiple files - only seems warranted for complex libs though
13:59:04	PMunch	If you compile the file you get the tool, but if you import it you won't get the tool included
14:01:05	WilhelmVonWeiner	that's pretty cool
14:01:22	WilhelmVonWeiner	I think I will get into the habit of writing proc main
14:03:27	FromDiscord	<mratsim> You can also do "when isMainModule: proc main() = "
14:06:13	Zevv	bwah. `converter fromFloat*[T, W](v: float): FixedPoint[T, W] =`: cannot instantiate `T`
14:07:47	*	liblq-dev joined #nim
14:09:15	FromDiscord	<mratsim> yep, need proc and typedesc (https://github.com/mratsim/trace-of-radiance/blob/master/trace_of_radiance/io/color_conversions.nim#L105-L106)
14:09:54	Zevv	yeah yeah but that sucks. I have my type for that
14:10:11	Zevv	my type `Speed` is `FixedPoint[uint8, 2]`
14:10:17	Zevv	and I just want to work with `Speed`
14:10:22	Zevv	and never have to refer to the definition
14:10:57	FromDiscord	<Recruit_main707> Generics can’t get their type from return types
14:11:28	FromDiscord	<Rika> you must explicitly define the generic types
14:11:34	FromDiscord	<Rika> since converters are implicit
14:11:41	Zevv	Yeah I know. but I hate it :)
14:11:44	FromDiscord	<Rika> its not really possible
14:12:22	Zevv	I'll make a `proc set[T, W](f: var FixedPoint[T, W], val: float)` then
14:15:02	Zevv	but I don't want to have floats at run time
14:15:06	Zevv	that should all be magiced away
14:25:11	*	narimiran quit (Ping timeout: 256 seconds)
14:25:49	FromDiscord	<mratsim> > its not really possible↵@Rika You never tried to break the law of causality. parameter (cause) -> return type (consequence) is such an outdated concept.
14:26:11	FromDiscord	<ache of head> 🤣
14:26:16	FromDiscord	<ache of head> Or…?
14:26:19	FromDiscord	<ache of head> Was this serious?
14:27:08	FromDiscord	<ache of head> My sarcasm and humor sensors are pretty bad compared to other people.
14:27:24	FromDiscord	<mratsim> Depends on the field, seems like in geopolitics and sociology, causality is walking on its head.
14:27:28	planetis[m]	you can try my fixed point lib https://gist.github.com/planetis-m/c13529cac44b4a9de3f073de502315ee
14:28:00	planetis[m]	there is another one in gitlab https://gitlab.com/lbartoletti/fpn
14:28:15	FromDiscord	<mratsim> I'm surprised you only have 16 gists @planetis, it feels like you have gist for everything to me
14:29:13	planetis[m]	well not everything, just simple stuff
14:29:28	PMunch	That looks promising, but I need to add a 6.2 to a 14.2
14:30:30	Zevv	PMunch: http://ix.io/2vct
14:30:39	planetis[m]	multithreading ie seems like rocket science to me
14:30:46	Zevv	it's a start, but it's cumbersome to make good mix & match possible between types
14:31:04	Zevv	this allows adding types with the same precision, assuming the first is larger or the same then the second
14:31:10	Zevv	so you can do Pos += Speed
14:31:45	FromDiscord	<mratsim> @planetis: start by a threadpool (which I didn't do :D)
14:31:55	FromDiscord	<mratsim> you can go far with a threadpool
14:32:05	FromGitter	<alehander92> zevv
14:32:09	FromGitter	<alehander92> i started with my ir/assembly stuff
14:32:42	FromGitter	<alehander92> can you give an example of an useful small thing to target: i want to target x86_64, maybe raspberry and 1 more
14:33:00	FromGitter	<alehander92> to have different things to make sure it's more general
14:36:47	FromDiscord	<mratsim> What do you want to use your IR / assembly for? Are you writing a compiler/code generator?
14:39:37	Zevv	yeah i don't know. he says its for "fun"
14:39:40	Zevv	"fun", he says
14:40:02	FromDiscord	<mratsim> I have "fun" writing bigints and crypto in pure assembly
14:40:08	Zevv	alehander: if you want to go extreme, target at least one 64 bit and one 8 bit device.
14:40:17	FromDiscord	<mratsim> I have even more "fun" debugging carry bugs/off-by-one
14:40:59	FromGitter	<alehander92> mratsim yeah a small subset of c one
14:41:08	FromDiscord	<mratsim> It's actually one area where the current compilers are broken
14:41:30	Zevv	alehander92: but making it 'generic' is not easy - different cpu's have different address modes, different numbers of registers, etc
14:41:36	FromGitter	<alehander92> mostly wanting to improve syntax errors / learn more about optimizations / start going deeper into c/asm for cv
14:41:40	FromDiscord	<mratsim> so, you can start with bigint addition on fixed sized bigint, sayd uint256
14:41:40	Zevv	so I'd be surprised if you could whip up something that would just work on x86 and arm
14:41:53	FromGitter	<alehander92> zevv yeah but this is good
14:41:56	FromDiscord	<mratsim> and then multiplication
14:42:23	FromGitter	<alehander92> that's the point, to make me think of what is abstract optimization and what is arch-specific
14:42:42	FromGitter	<alehander92> well, 8bit .. might be a bit strange
14:42:56	FromGitter	<alehander92> how easy can i find such a device/emulator, chip8?
14:43:12	Zevv	pick avr, so much available for that
14:43:18	Zevv	they call it 'arduino' these days
14:43:43	FromGitter	<alehander92> what other things are "strange"
14:43:57	FromGitter	<alehander92> are there common cpus without registers
14:43:59	FromGitter	<alehander92> or without heap
14:44:19	FromGitter	<alehander92> (ok the last thing is probably userland-specific)
14:44:40	FromGitter	<alehander92> ok, with 4bit words or 10bit
14:44:44	FromGitter	<alehander92> or something like that
14:44:48	Zevv	don't exist anymore these days
14:45:02	Zevv	but really man, just target 1 first
14:45:05	Zevv	that's hard enough as it is
14:45:19	FromGitter	<alehander92> yeah, but i want 2
14:45:21	FromGitter	<alehander92> otherwise my ir
14:45:25	FromGitter	<alehander92> would fit the 1 too much
14:46:10	Zevv	arm thumb
14:49:52	FromDiscord	<mratsim> 12-bit and 24-bit are likely to be audio DSP
14:49:54	*	hnOsmium0001 joined #nim
14:50:24	FromDiscord	<mratsim> AVR has a lot of crypto/bigint implementation
14:50:38	FromDiscord	<mratsim> so big int multiplication on it is interesting
14:50:51	FromDiscord	<mratsim> also ARM book as a big int in assembly section
14:50:57	Zevv	well, interesting. it's just carrying and carrying and carrying
14:51:08	Zevv	"do stuff lots of times"
14:51:52	FromDiscord	<mratsim> https://media.discordapp.net/attachments/371759389889003532/747830605236142141/unknown.png
14:52:06	FromGitter	<alehander92> hm, i am more interested in non-math opt
14:52:26	FromGitter	<alehander92> math ones are very cool, but :D I mostly wanted to go through gcc flags
14:52:30	FromGitter	<alehander92> and see what is easy to learn
14:52:59	Zevv	PMunch: ping
14:53:02	FromGitter	<alehander92> i mostly want something that i can easily tst on
14:53:05	PMunch	Pong
14:53:08	FromDiscord	<mratsim> My advice, pick something that you can show
14:53:16	FromDiscord	<mratsim> like raytracing *cough*
14:53:21	FromGitter	<alehander92> so i should either get it from a friend/buy or find an emulator
14:53:21	Zevv	PMunch: that ix I just posted
14:53:35	PMunch	Oooh
14:53:36	Zevv	can you think of a way to make the `set()` compile time
14:53:38	FromGitter	<alehander92> mratsim haha, well i prefer parsing
14:54:01	disruptek	alehander92: pick some gameboy stuff or whatever, then you'll have lots of roms to test with.
14:54:04	FromDiscord	<mratsim> parse a demo scene file and transform it to many arch
14:54:13	FromDiscord	<mratsim> or a mandelbrot fractal
14:54:26	FromGitter	<alehander92> i should look at common benchmarks
14:54:41	FromGitter	<alehander92> but i am far away from this
14:54:56	FromDiscord	<mratsim> Are compile brainfuck to multiple arch
14:55:04	PMunch	Make a toFixed compileTime procedure and make set a template that expands to "f = toFixed[T, W](val)"?
14:55:10	FromGitter	<alehander92> no i dislike it
14:55:59	FromGitter	<alehander92> but thanks man
14:56:08	FromGitter	<alehander92> i should keep in mind some math usecases
14:56:23	FromGitter	<alehander92> but this is not really a super serious project
14:57:10	FromGitter	<alehander92> disruptek thanks
14:57:44	Zevv	PMunch: but I want to be able to make arrays and seqs as well, without havinc to specify for each element
14:57:56	FromGitter	<alehander92> i think i might target rpi / arduino / something very old like chip8 just for ease of test first
14:57:57	PMunch	Ah..
14:58:09	FromGitter	<alehander92> (and x86_64) later more, thanks for the tips
14:58:14	PMunch	And I guess converters can't be {.compileTime.}?
14:58:24	Zevv	I can't even convert this because the type is generic
15:00:06	*	MarderIII joined #nim
15:02:50	Zevv	bah that would have been nice if converters worked
15:03:18	disruptek	when you have the money you need, it doesn't matter how productive you are because what is production at that point?
15:04:07	disruptek	i started three different businesses, and was busy inventing shit all the time. but for what?
15:04:21	Zevv	well, did you do it for the money?
15:04:50	Zevv	I asked my accountant once what I would need to do to get rich. "Well, not what you're doing now" was the answer
15:05:20	disruptek	that's the point; you have to look at this from the other perspective.
15:05:25	disruptek	it's like star trek.
15:05:40	disruptek	once you can just hit a button and get food, you need new goals.
15:05:56	Zevv	and once you're the ADHD entrepreneur type of person, you need new goals
15:06:05	Zevv	all the time
15:09:25	disruptek	if you combine that person with a button that doesn't produce food, you have a real problem.
15:10:11	Zevv	what's the name of the function that should be called "sign" that does if v < 0.0: -1.0 else: 1.0
15:10:33	Zevv	oooh. `sgn`
15:11:03	disruptek	just use ndoc
15:11:17	disruptek	`ndoc sign` it's the second hit.
15:11:36	Zevv	ndoc
15:11:37	Zevv	wut
15:11:50	disruptek	alehander92's new project.
15:11:53	disruptek	!repo disruptek/ndoc
15:11:54	disbot	https://github.com/disruptek/ndoc -- 9ndoc: 11like pydoc but for nim 15 0⭐ 0🍴
15:12:00	Zevv	dude
15:12:03	Zevv	talking about productivity
15:12:26	FromDiscord	<ache of head> Hey, what was the operator for explicitly indicating types of number literals? Like, I have a 1 that I want to be a uint8. I think there was something like an apostrophe…?
15:12:26	Zevv	you're not the ADHD entrepreneur type, are you
15:12:45	*	kenran quit (Ping timeout: 240 seconds)
15:12:46	disruptek	1.uint8 or 1'u8
15:12:56	Zevv	or uint8(1)
15:12:57	FromDiscord	<ache of head> thank you very much!
15:12:57	*	Vladar joined #nim
15:13:24	disruptek	1'u8 is annotation, the others are conversion.
15:13:44	FromDiscord	<ache of head> I see
15:13:44	FromDiscord	<ache of head> Thanks!
15:14:27	disruptek	what i've learned about ADHD is that it manifests in mysterious ways.
15:15:10	Zevv	it does, it does
15:17:50	FromDiscord	<Rebel> Let's say hypothetically someone wanted to do inbranch llvm obfuscation such as with YANSOllvm↵how hard would that be to do with Nim? Would the easiest way just be using clang with the c files↵that are produced?
15:18:00	Zevv	eh disruptek you're smart. think for me.
15:18:11	Zevv	https://github.com/zevv/nimfix/blob/master/fixedpoint.nim
15:18:36	Zevv	I have a type `FixedPoint[T, W]`. to be used like `type Speed = FixedPoint[uint8, 2]`
15:18:56	FromDiscord	<Rebel> (edit) 'inbranch' => 'in tree'
15:19:13	Zevv	I want to be able to say something like "var s: Speed = 1.25" and `var s: seq[Speed] = @[ 1,25, 3.5 ]`
15:19:13	disruptek	rebel: yes, just use clang. it works fine.
15:19:22	Zevv	but converters don't fly beause of the generics it seems
15:19:40	Zevv	how would you make this so that its both consice, and that the floats don't survive to run time
15:19:53	disruptek	you can run the converter to instantiate the generics, right?
15:20:00	disruptek	prime the pump.
15:20:01	Zevv	converter says no.
15:20:05	Zevv	can't innstatitate T
15:20:28	disruptek	are you sure you know what i'm saying?
15:20:43	Zevv	I have no clue
15:20:46	Zevv	really
15:20:57	FromDiscord	<mratsim> welcome to low level programming "I have no clue" 😛
15:21:25	Zevv	I hardly ever do
15:21:30	Zevv	it's mostly talk
15:21:40	Zevv	I got a way for that for many years
15:22:05	Zevv	and the fun part is: the further I get in my career, the more often I say "dude I have no clue what that is". and people kind of respect that. really.
15:23:56	disruptek	so the way generics work is, basically, once they've been called with types, they can be called without types.
15:24:20	Zevv	sure. but how does one "prime the pump"
15:24:38	Zevv	I have ideas, but these dont apply to nim
15:25:56	disruptek	put the values in the box[](somevar)
15:26:50	disruptek	conv[int8, 3](dx) looks like
15:30:37	Zevv	hmm first I shell cook dinner, so I have been told
15:30:58	disruptek	we need a trampoline that lets you register into different stages.
15:31:15	disruptek	so you can catch continuations pre-bounce, post-bounce.
15:32:08	Zevv	why "need"
15:32:12	Zevv	i think we need tons of stff
15:32:43	disruptek	i mean, it needs to come in the box.
15:34:01	disruptek	this is basically what our tubes are built on, but it's so generic a concept that it should be available outside of tubes.
15:34:45	disruptek	i have to get coffee and shoot a horse, but i'll look at your converter when i get back.
15:35:02	FromDiscord	<ache of head> Oof
15:35:09	FromDiscord	<ache of head> I figured that the static seq doesn't work
15:35:11	FromDiscord	<ache of head> cause it's not static xD
15:52:06	*	dddddd quit (Remote host closed the connection)
15:53:46	*	dddddd joined #nim
15:58:24	Zevv	yeah sure butem in the box, but i dont want to dry
15:58:40	Zevv	theres a type like Speed that is already [uint8, 2]
15:58:50	Zevv	hm this smells so gooood
15:59:10	disruptek	this is bullshit.
15:59:44	Zevv	its hard to put pig in things and make'm smell bad
16:00:03	disruptek	i beg to differ.
16:00:15	disruptek	witness aged cream of bacon soup.
16:00:35	disruptek	"it's raining too hard to shoot horses"
16:00:42	disruptek	how is this even a fucking thing?
16:00:47	Zevv	it si, it is, so says the horse
16:00:53	disruptek	damnit vermont sucks.
16:01:30	Zevv	try vermouth
16:01:59	FromDiscord	<lqdev> Zevv: https://github.com/zevv/nimfix/blob/master/fixedpoint.nim#L40
16:01:59	FromDiscord	<lqdev> lol
16:02:11	disruptek	okay, i ran your file.
16:02:13	*	dsrw joined #nim
16:02:15	disruptek	what am i supposed to get?
16:02:36	*	liblq-dev quit (Quit: WeeChat 2.9)
16:06:09	disruptek	your converter converts the wrong way.
16:08:38	Zevv	no this is the other way
16:08:44	Zevv	the right one is not in here yet
16:09:22	Zevv	pull, line #13
16:10:47	Zevv	its not the usage that is erroring, its the definition of the converter itself
16:11:43	disruptek	i guess my last lady visitor drank up the vermouth.
16:12:03	Zevv	next time, shoot her, rain or not
16:12:34	Zevv	but what about the converter. what do i put in whose box
16:12:48	Zevv	because I tried putting thing in boxes
16:12:52	Zevv	but that didn't work out
16:13:25	Zevv	these are pretty nasty generics like this
16:13:31	Zevv	[T1, T2, W1, W2]
16:13:32	Zevv	brr
16:14:26	*	dsrw quit (Quit: Textual IRC Client: www.textualapp.com)
16:14:43	*	dsrw joined #nim
16:14:54	disruptek	i dunno how this can work.
16:15:07	Zevv	but I still respect you
16:15:15	disruptek	weird.
16:15:22	Zevv	yes, I should loathe you
16:15:37	disruptek	well, let's not go overboard.
16:15:58	Zevv	but as always the question arises: is this a bug, and/or am I plain stupid?
16:16:28	Zevv	the prefered answer on #nim is "generics cant instantiate on return type"
16:16:50	disruptek	nah.
16:17:03	disruptek	this isn't that. you aren't providing types anywhere.
16:17:16	*	waleee-cl quit (Quit: Connection closed for inactivity)
16:17:23	disruptek	it's like an equation it cannot solve because there are too many variables.
16:17:29	Zevv	sure. I know that.
16:17:41	disruptek	make it work as procs, first. i'm wary of the converter interaction.
16:17:47	Zevv	Thus came my original question: would you know some way to do this consise and neat
16:17:57	Zevv	I normally don't use converters.
16:18:01	Zevv	but when I do
16:18:03	Zevv	I do it for PMunch
16:18:17	disruptek	they have such limited use.
16:18:41	Zevv	My problem is that I have a type that is a "specified" FixedPoint type
16:18:48	Zevv	like Speed being a FixedPoint[int8, 2]
16:18:57	Zevv	Now I want to assign a variable of type Speed with a float
16:19:03	Zevv	and this float should only exist a compile time
16:19:12	Zevv	and even better, I want to do that with arrays or seqs as well
16:19:23	Zevv	that was my original intent. The converters came in as a side effect when I wasn't looking
16:19:32	disruptek	yeah, it doesn't work as a converter but it works as a proc.
16:20:07	Zevv	yeah, but then you have to repeat your type. Like 'uint8, 2'
16:20:11	Zevv	but I want to talk about Speed
16:20:58	disruptek	seems to be due to interaction with other stuff.
16:21:07	Zevv	it's like life
16:21:19	disruptek	messy.
16:21:23	Zevv	messy.
16:21:54	disruptek	i tried to get into a house last night and there was a huge spider on the handle of the door.
16:22:01	disruptek	i was like, "not tonight, motherfucker!"
16:22:08	*	snowolf quit (Remote host closed the connection)
16:22:26	Zevv	and who's house was that
16:22:30	*	dsrw_ joined #nim
16:22:30	disruptek	knocked him off and go inside. then when i went to leave, i slipped on the steps and fell on my ass.
16:22:39	disruptek	the door hit me in the face twice.
16:22:46	Zevv	and the spider?
16:22:51	disruptek	if they don't get me comin', they get me goin'.
16:23:01	disruptek	dunno what happened to the spider.
16:23:03	Zevv	messy.
16:23:06	disruptek	messy.
16:23:12	*	dsrw quit (Quit: Textual IRC Client: www.textualapp.com)
16:23:16	*	snowolf joined #nim
16:23:25	Zevv	ok, I'll see if I can macro this somehow
16:23:33	disruptek	holdon, i'm not done with it.
16:23:41	Zevv	neither am I
16:24:37	disruptek	its fundamentally never going to work because it has no idea what you're trying to create.
16:25:25	Zevv	well, the converter ist
16:25:32	Zevv	what I want to do is basicalyl this:
16:25:34	disruptek	you need a converter that converts to Speed.
16:25:35	Zevv	var a: Speed = 3.1415
16:25:39	*	dsrw_ is now known as dsrw
16:25:45	Zevv	yes, but I don't want to make a converter for every thing
16:25:46	disruptek	or, make Speed.set.
16:25:48	Zevv	Nim knows what Speed is
16:25:58	Zevv	Speed is a FixedPoint[int8, 2]
16:26:06	Zevv	so I don't want to have to say that twice
16:26:10	Zevv	I have kids for that
16:26:40	disruptek	where do you need the converter?
16:26:55	Zevv	var dx: Speed = -1.3
16:27:04	Zevv	Initialization stuff. I want to talk human floats in my source
16:27:12	disruptek	ahhh, okay, lemme see.
16:27:36	Zevv	or var ss: seq[Speed] = @[ -1.3, 2.4 ]
16:29:22	disruptek	seems like this should be possible, then.
16:30:11	FromGitter	<cloudhan> `var a = 1 -2` triggers a compiling error, is the space significant or it is a compiler bug?
16:30:20	disruptek	stupid return value dispatch.
16:31:19	disruptek	cloudhan: - is a unary operator.
16:31:31	Zevv	space is significant. This parses as `var a = (1) (-2)`
16:31:38	Zevv	it's deliberate
16:33:40	disruptek	the irony is that converters are by their very nature "dispatch on return type."
16:34:31	Zevv	until they don't
16:34:49	Zevv	they are like generic methods. The work. Sometimes.
16:35:22	disruptek	we need to prime the pump.
16:37:46	disruptek	#2869
16:37:48	disbot	https://github.com/nim-lang/Nim/issues/2869 -- 3Generic converter with no generic param doesn't compile when there is another generic converter with the same name ; snippet at 12https://play.nim-lang.org/#ix=2vd8
16:41:43	Zevv	not sure if this is my problem, is it?
16:42:00	disruptek	no, just funny coincidence.
16:42:00	Zevv	hm it is
16:42:05	disruptek	nah.
16:42:13	Zevv	n't
16:42:24	Zevv	i wasn't done typing
16:42:27	Zevv	have patience
16:42:36	disruptek	lol
16:43:33	disruptek	#4554
16:43:34	disbot	https://github.com/nim-lang/Nim/issues/4554 -- 3converter fails to work with multiple generic types
16:44:44	Zevv	"showstopper"
16:44:47	Zevv	ha. ha.
16:45:08	disruptek	zah is working on it.
16:45:16	Zevv	since 2017
16:45:23	disruptek	these things take time.
16:45:30	disruptek	you want it done right, don't you?
16:46:56	disruptek	some day i'm going to have to learn the frontend.
16:49:47	FromDiscord	<Rika> how do i increase the buffer size of stdio (and is this even a good idea?)a
16:49:51	FromDiscord	<Rika> (edit) 'idea?)a' => 'idea?)'
16:50:18	leorize[m]	why would you want to increase the buffer size?
16:51:42	FromDiscord	<Rika> i feel like the flickering when i write to the same line is because the buffer is flushed out before i manually flush it
16:53:16	disruptek	are you using a single write?
16:53:42	FromDiscord	<Rika> no
16:53:44	disruptek	Zevv: use a macro.
16:53:48	FromDiscord	<Rika> let me try that
16:55:16	FromDiscord	<Rika> seems to be better yeah
17:00:16	*	vsantana joined #nim
17:04:30	*	vsantana quit (Client Quit)
17:08:29	Oddmonger	does «const» belong to macros ?
17:08:40	Oddmonger	i think so, but not sure
17:09:12	dsrw	I'm trying to better understand how procs returning `var string` are supposed to work, specifically https://github.com/nim-lang/Nim/blob/devel/lib/pure/parsecsv.nim#L336. It seems like a `var string` proc doesn't have to return anything, but, if it doesn't, there's no way to check for `nil`, so it's basically a guaranteed SIGSEGV.
17:09:14	*	kenran joined #nim
17:10:59	dsrw	Do I have that right? Is this behaviour something that may change in the future?
17:11:47	FromDiscord	<Rika> wdym? strings arent nilable?
17:12:22	FromDiscord	<Rika> ah wait
17:12:33	leorize[m]	dsrw: that example is two years old
17:12:56	leorize[m]	I'm pretty sure a lot have changed since :P
17:13:31	dsrw	Agreed, but in this case you still get the SIGSEGV, at least on 1.2.6.
17:14:05	*	kenran quit (Ping timeout: 240 seconds)
17:14:24	leorize[m]	ah yea, the compiler should enforce a return value for `var T`
17:14:56	leorize[m]	the fix for that module in particular is to just throw a KeyError exception
17:16:00	dsrw	Ok. That's what I thought. I'll check github for a similar issue, and file one if it doesn't exist. Thanks!
17:16:27	dsrw	(and maybe submit a fix for parsecsv while I'm at it)
17:18:27	disruptek	there are a few cases like that in stdlib.
17:20:09	Zevv	so, I use a macro, right
17:20:12	Zevv	the pain
17:20:13	Zevv	the agony
17:20:51	dsrw	Any thoughts on why that proc is using `var string` to begin with?
17:22:44	dsrw	It seems like `string` would be fine. I can't think of why you'd need to mutate the result.
17:24:20	PMunch	Zevv, aaw you're even using converters for me :) I guess I have to seriously consider your marriage proposal :P
17:24:34	Zevv	nope, converters are out. broken, so broken
17:24:41	Zevv	I now need to do macros with types
17:24:44	Zevv	the things I do for love
17:25:00	Zevv	How the *Hell* do I get to the generics passed to my FixedPoint from inside the macro
17:25:16	Zevv	Usually I try all permutations of getType, getTypeImpl and getTypeInst until I get something that looks ok
17:25:17	FromDiscord	<Clyybber> Zevv: Whats the W generic param for?
17:25:19	Zevv	but it's not coming out today
17:25:31	Zevv	Bits after the comma
17:25:32	FromDiscord	<Clyybber> oh, I see it now
17:25:34	FromDiscord	<Clyybber> yeah
17:25:35	Zevv	static(int)
17:25:56	Zevv	making fixed point libs with fixed sizes is for sissies
17:25:59	Zevv	real men do this fully generic
17:26:07	FromDiscord	<Clyybber> yeah
17:26:09	Zevv	and run into compiler bugs on the road
17:26:18	Zevv	then we summon our knight
17:26:23	Zevv	may he slay these monsters for us
17:26:29	FromDiscord	<Clyybber> and scream NIM NIM NIM
17:26:47	Zevv	no, we just *say* that
17:27:11	Zevv	so are there more getTypeThings I can do
17:27:20	PMunch	Haha, I seriously need to order one of those tshirts
17:27:23	Zevv	I know my S is a FixedPoint
17:27:28	Zevv	but I don't know what's in the box
17:27:29	PMunch	I wonder if FOSDEM 2021 will happen..
17:27:36	PMunch	Oh well, got to go
17:27:38	*	PMunch quit (Quit: leaving)
17:27:45	Zevv	PMunch: until we meet again!
17:28:47	disruptek	stupid concepts.
17:29:52	Zevv	stupid macros
17:29:57	Zevv	can't get to my genericParams
17:32:21	Zevv	How the *hell* was I supposed to know that there is a `symbol` field in my getTypeInst[1] that I can do getImpl on
17:32:25	Zevv	how. how?!
17:47:39	disruptek	leorize[m]: i'm attempting to adapt your ci. naively edited the file and pushed it.
17:48:13	disruptek	far superior to travis so far.
17:48:31	disruptek	if it's really this easy, i will integrate it into nimph asap.
17:54:56	disruptek	leorize[m]: if this stuff only looks at changes to src/ or tests/ then that's a problem; it should track .cfg, .nimble, .nims files, too, and any .nim that isn't in src/ as well.
17:54:57	*	narimiran joined #nim
17:57:05	Zevv	https://github.com/nim-lang/Nim/issues/15225
17:57:07	disbot	➥ Generic converter: cannot instantiate ; snippet at 12https://play.nim-lang.org/#ix=2vdA
17:57:29	Zevv	I must just be doing something wrong. This is a very simple thing.
17:57:41	Zevv	disruptek: what's "leorize's ic" ?
17:58:01	*	bra-ket joined #nim
17:58:23	disruptek	that repro crashes dust.
17:58:52	FromDiscord	<Clyybber> Zevv: Why would that work
17:58:54	disruptek	leorize has been developing github actions for nim test matrices.
17:58:57	FromDiscord	<Clyybber> Your T is floating in the air
17:59:17	*	zyklon quit (Ping timeout: 260 seconds)
17:59:23	Zevv	clyybber: I must be stupid today. 's' has a type, right?
17:59:27	disruptek	because it's typed.
17:59:27	Zevv	T is 'uint8' for s
17:59:27	FromDiscord	<Clyybber> because nim doesn't have return type overloading
17:59:37	FromDiscord	<Clyybber> yeah, but nim doesn't have that feature
17:59:42	Zevv	but isn't that the whole thing for converters?
17:59:42	FromDiscord	<Clyybber> *yet*
17:59:49	Zevv	isn't that what converters *are*
17:59:52	FromDiscord	<Clyybber> fair point
17:59:55	FromDiscord	<Clyybber> lol
18:00:53	disruptek	the irony is that converters are by their very nature "dispatch on return type."
18:00:57	Zevv	so, should i delete the issue then?
18:01:02	disruptek	no.
18:01:10	Zevv	but now everybody will think I am stupid
18:01:14	Zevv	s/think/know/
18:01:16	disruptek	nah.
18:01:18	Zevv	we can't have that
18:01:43	disruptek	hey i fixed my stupid macro exception issue.
18:01:59	Zevv	which one of your stupid macro exception issues is that
18:02:03	FromDiscord	<Clyybber> nah everybody will say, you stupid
18:02:08	FromDiscord	<Clyybber> and then realize me stupid
18:02:18	Zevv	and then realize clyybber stupid?
18:02:26	FromDiscord	<Clyybber> disruptek: Fixed as in fixed in nim?
18:02:30	FromDiscord	<Clyybber> Zevv: ye
18:02:36	Zevv	hehe
18:03:07	disruptek	nah.
18:03:31	disruptek	#15160
18:03:33	disbot	https://github.com/nim-lang/Nim/issues/15160 -- 5exceptions codegen broken inside macros ; snippet at 12https://play.nim-lang.org/#ix=2vdJ
18:03:52	disruptek	someone might decide that it's still a bug, but i'm okay with separate symbols.
18:04:10	FromDiscord	<Clyybber> Can you explain what symbols you mean?
18:04:21	disruptek	except Foo as bar: # symbol bar
18:04:37	FromDiscord	<Clyybber> and you can't have an except Bar as bar after that?
18:04:42	disruptek	right.
18:04:45	FromDiscord	<Clyybber> Thats a bug
18:04:46	Zevv	that's funny
18:04:49	disruptek	not with the same gensym'd bar.
18:04:57	FromDiscord	<Clyybber> well
18:04:58	Zevv	it's a different scope, isn't it?
18:05:01	disruptek	it is.
18:05:02	FromDiscord	<Clyybber> hmm
18:05:10	*	myphs joined #nim
18:05:16	FromDiscord	<Clyybber> you using the same sym is weird for sure
18:05:34	FromDiscord	<Clyybber> so yeah, not a bug, but could have a nicer error message I suppose
18:05:38	*	vsantana joined #nim
18:06:03	disruptek	if you could do that, you could probably solve the bug.
18:06:24	Zevv	hehe
18:06:24	*	kungtotte quit (Read error: Connection reset by peer)
18:09:53	*	kungtotte joined #nim
18:11:04	leorize[m]	"disruptek" (https://matrix.to/#/@freenode_disruptek:matrix.org): I only track those folder due to how my project structure work
18:11:27	disruptek	yep.
18:11:36	leorize[m]	you can use a generic "on: [push, pull_request]" and it will run on those events unconditionally
18:12:08	Zevv	disruptek: https://axelforsman.tk/2020/08/24/rust-style-futures-in-c.html
18:12:21	disruptek	yeah, i haven't read it yet.
18:12:26	Zevv	neither have I :)
18:12:40	FromDiscord	<Clyybber> wait, I'm not crazy right. defined and declared is not the same
18:12:49	*	hyiltiz quit (Ping timeout: 264 seconds)
18:13:17	disruptek	i dunno what declared is.
18:13:31	FromDiscord	<Clyybber> it checks if a symbol is declared
18:13:38	FromDiscord	<Clyybber> and defined checks if its defined
18:13:54	FromDiscord	<Clyybber> via -d and the like
18:13:59	FromDiscord	<Clyybber> the weird thing is
18:14:02	FromDiscord	<Clyybber> they share a magic
18:14:09	disruptek	lol
18:14:33	disruptek	maybe declared works on defines.
18:14:42	*	Zectbumo joined #nim
18:14:44	Zevv	ok I read it. it sucks
18:14:49	disruptek	Zevv: that's the kind of article dom96 will like.
18:14:53	FromDiscord	<Clyybber> not really, their implementation simply branches on the name of your invocation
18:15:02	disruptek	he loves to use other language designs as justification.
18:15:05	Zevv	it's duffs fake coroutines with futures
18:15:15	*	tane joined #nim
18:15:29	Zevv	the do the duffs thing because they must in C. Rust knows how to do the cps
18:15:34	Zevv	"turn the code into a state machine"
18:15:37	Zevv	hell yeah
18:15:58	Zevv	once you did that properly, you can drop the futures
18:16:19	Zevv	"The coroutines for implementing the future state machine we have described are stackless,"
18:16:44	Zevv	"CPS is just about moving the stack to the heap" - Zevv
18:16:55	disruptek	not just that.
18:17:03	disruptek	the stack becomes dynamic.
18:17:06	Zevv	anyway, PMunch, you can do that fixedpoint stuff yourself. I hate this
18:17:09	disruptek	you can resize it, retype it, etc.
18:17:19	FromDiscord	<Clyybber> Zevv: :(
18:17:22	Zevv	daft punk should make a song about that
18:17:27	FromDiscord	<Clyybber> lol
18:17:32	disruptek	i don't know how to do things like clone stacks without cps.
18:17:54	Zevv	clone where to? there is only one stack
18:17:58	Zevv	so yes
18:18:40	*	vicfred quit (Read error: Connection reset by peer)
18:18:47	*	vicfred_ joined #nim
18:30:29	Zevv	so now instead of doing smart things with generic converters
18:30:38	Zevv	I now create a template that defines the type and also makes a converter
18:30:46	FromDiscord	<Clyybber> why though
18:30:50	FromDiscord	<Clyybber> why all this stuff
18:30:56	Zevv	because of love
18:31:07	FromDiscord	<Clyybber> why not simply say make your conversions explicit
18:31:12	*	disruptek will eat for food.
18:31:14	FromDiscord	<Clyybber> nobody likes implicit conversions anyways
18:31:19	Zevv	Clyybber: it's *too* explicit
18:31:29	Zevv	No I don't need *converters* perse
18:31:35	Zevv	but I need functions to work with my 'friendly' types
18:31:37	*	vsantana quit (Read error: Connection reset by peer)
18:31:37	FromDiscord	<Clyybber> Zevv: what is
18:31:38	disruptek	i think he lost a bet with pmunch.
18:31:46	FromDiscord	<Clyybber> what is too explicit
18:31:46	*	vsantana joined #nim
18:31:50	Zevv	I have a type 'Speed' which is 'FixedPoint[int16, 4]'
18:31:55	Zevv	I want to be able to convert 3.5 to Speed
18:31:56	FromDiscord	<Clyybber> yeah
18:32:01	FromDiscord	<Clyybber> Speed(3.5)
18:32:06	FromDiscord	<Clyybber> is what should be possible
18:32:07	Zevv	No, that's wrong
18:32:19	*	andinus quit (Remote host closed the connection)
18:32:20	Zevv	speed is a distinct int type
18:32:24	Zevv	so that will just assign 3
18:32:41	FromDiscord	<Clyybber> overwrite the explicit conversion
18:32:47	Zevv	how
18:32:54	Zevv	`=` ?
18:33:05	FromDiscord	<Clyybber> define a proc with the same name
18:33:10	Zevv	dude
18:33:12	FromDiscord	<Clyybber> I'm not sure that will overwrite it tho
18:33:14	Zevv	really
18:33:14	disruptek	leorize[m]: why does this ci build nim instead of using your nightlies artifacts?
18:33:33	*	hyiltiz joined #nim
18:33:33	*	hyiltiz quit (Changing host)
18:33:33	*	hyiltiz joined #nim
18:33:47	Zevv	Clyybber: you're just making this up, are you
18:33:51	FromDiscord	<Clyybber> nah
18:34:28	Zevv	redefinition of 'Speed', previous declaration here
18:34:45	FromDiscord	<Clyybber> ah, damnit.
18:34:55	FromDiscord	<Clyybber> Then it must have worked because I had the original type not exported
18:35:03	FromDiscord	<Clyybber> well, I think we should make it work
18:35:06	FromDiscord	<Clyybber> make it an idiom
18:35:09	disruptek	honestly, this is the sort of abstraction that we need to get right.
18:35:21	disruptek	it has to be easier than this to hide stuff.
18:35:34	Zevv	but then again: `Speed` is a derived type. I don't want to create a proc to make a Speed every time I define the type
18:35:44	Zevv	the user makes his type, and then should be able to work with it
18:35:54	Zevv	you can't ask the user to define the type as a specific FixedPoint
18:36:00	Zevv	and then also create a proc to do the conversion
18:36:01	disruptek	point made, zevv, just need to decide which bugs you must fix to achieve it.
18:36:23	Zevv	just let me make my point, damnit!
18:37:45	*	hyiltiz quit (Ping timeout: 240 seconds)
18:40:22	leorize[m]	disruptek: scripting to get the latest artifact is a PITA, not to mention that my new nightlies still hasn't been merged
18:40:37	leorize[m]	I've an idea on how to solve the former, so there's that
18:40:46	disruptek	hmm, lemme know if i can write something.
18:41:23	disruptek	this stuff should be much easier once king's graphql is ready.
18:42:27	*	hyiltiz joined #nim
18:42:27	*	hyiltiz quit (Changing host)
18:42:27	*	hyiltiz joined #nim
18:43:42	*	Jesin quit (Quit: Leaving)
18:43:44	*	pbb quit (Quit: http://quassel-irc.org - Chat comfortably. Anywhere.)
18:44:17	*	pbb joined #nim
18:49:40	*	pbb quit (Remote host closed the connection)
18:50:54	*	pbb joined #nim
18:51:10	*	Jesin joined #nim
18:51:45	*	Zectbumo quit (Read error: Connection reset by peer)
18:55:38	*	a_b_m joined #nim
18:55:56	*	waleee-cl joined #nim
18:57:42	FromGitter	<kingramx> I searched whole the internet including deepweb and darkweb :/ and couldn't find anything about how retrieve json-data from .json file and use them in .nim file ⏎ could someone please help ???
18:59:06	*	abm quit (Ping timeout: 260 seconds)
19:00:17	krux02	disruptek, I have a question about math.round, why was it undeprecated?
19:00:27	Zevv	I'm pretty sure the darkweb should have that info kingramx
19:00:31	FromDiscord	<özgürlük> I'd like to ask a question, is there a library for making discord bots in nim?
19:00:39	disruptek	did i undeprecate math.round?
19:00:50	FromDiscord	<Rosen> json parsing is a dark art
19:00:56	krux02	oh sorry, you did not it was narimiran
19:00:59	FromDiscord	<Rika> @özgürlük `dimscord`
19:01:03	krux02	not you
19:01:05	disruptek	okay.
19:01:24	krux02	narimiran, why did you undeprecate math.round?
19:01:32	krux02	https://github.com/nim-lang/Nim/pull/15224
19:01:33	disbot	➥ remove deprecation from `math.round`
19:01:42	Zevv	kingramx: did you also happen to read the 'json' library page of the nim stdlib documentation?
19:02:06	FromDiscord	<özgürlük> @Rika thanks, the discord library i searched was super outdated
19:02:20	narimiran	krux02: yes, i did it, because we decided it is the best to de-deprecate it
19:02:29	krux02	best for what?
19:02:45	krux02	I deprecated it, because it should be removed without substitution.
19:02:51	narimiran	IYO
19:02:59	krux02	?
19:03:34	krux02	what does IYO mean?
19:03:38	narimiran	in your opinion
19:03:54	FromDiscord	<Clyybber> If we deprecate it we should replace it with quantize
19:04:05	disruptek	lol
19:04:07	FromDiscord	<Clyybber> but not this weird "drop your round in" implementation of quantize
19:04:11	krux02	why replace it with anything?
19:04:24	FromDiscord	<Clyybber> because a stdlib should provide stuff you often need
19:04:30	FromDiscord	<Rika> because many programmers use round?
19:04:34	FromGitter	<alehander92> krux02 there is a quantize pr
19:04:35	FromDiscord	<Rika> and think of using round?
19:04:40	FromGitter	<alehander92> maybe he's not aware of it
19:04:48	FromDiscord	<Clyybber> alehander92: That one was closed, he's talking about that pr
19:04:55	krux02	I argued that decimal rounding in a math library for binary numbers doesn't make sense and serves no purpose.
19:04:56	FromGitter	<alehander92> ah sorry
19:05:11	FromGitter	<kingramx> @Zevv yeah, I read that doc but u know document does not contain anything about how to retrieve data from json file and only talks about how to parse our data into json
19:05:45	narimiran	"serves no purpose" -- citation needed
19:05:45	krux02	decimal rounding does make sense if you have a decimal floating point number type (they do exist)
19:05:48	FromDiscord	<Rika> the math library works on binary-represented numbers but they process in decimal
19:06:11	disruptek	kingramx: let js = parseJson(readFile("somefilename.json"))
19:07:29	krux02	narimiran, "decimal rounding [...] serves no purpose", krux02:21:04:55
19:07:38	krux02	there, I cited myself :P
19:07:44	disruptek	makes sense to me.
19:07:53	disruptek	i mean, krux02's argument.
19:08:01	narimiran	great, now i'm going to enjoy my evening. ping me tomorrow if you want to argue some more
19:08:54	krux02	narimiran, I read the issue that is linked to the PR that you made, but I don't see the use case for the rounding function that you want to preserve.
19:09:10	krux02	The only use case originally was for printing and there it didn't work reliably.
19:09:17	*	quetzalb quit (Remote host closed the connection)
19:09:26	krux02	therefore there was the message to use strutils instead
19:09:35	krux02	the message that you gracefully removed.
19:10:04	*	kenran joined #nim
19:10:26	disruptek	narimiran: is it a bad idea to backport exitprocs?
19:10:55	krux02	narimiran, did you leave?
19:11:00	narimiran	disruptek: link?
19:11:03	narimiran	krux02: trying to
19:11:06	FromDiscord	<Lod> clear
19:11:07	FromGitter	<kingramx> @disruptek Really thanks for this
19:11:23	disruptek	narimiran: it's a module in std?
19:11:25	*	pbb quit (Quit: http://quassel-irc.org - Chat comfortably. Anywhere.)
19:11:39	FromDiscord	<haxscramper> When I use `{.byaddr.}` in macro I get `Error: invalid pragma: byaddr` error. Example: https://play.nim-lang.org/#ix=2ve5 . What can cause this?
19:11:55	FromDiscord	<haxscramper> Using it for regular variables works as expected
19:11:59	krux02	narimiran, I am trying to get an explanation from you.
19:12:16	krux02	I would really like to know about the use case for the deprecated round function.
19:12:31	disruptek	kingramx: hope it helps.
19:13:01	narimiran	disruptek: i'll see tomorrow about it
19:13:01	krux02	I can only think of use cases where `round` would be an false friend that provides you with something that doesn't work
19:13:25	disruptek	narimiran: i have no opinion. just noticed tests failing. 😉
19:13:50	krux02	and that is why I wanted the function to be gone.
19:14:18	*	pbb joined #nim
19:14:34	disruptek	krux02: maybe it's enough just to provide a compile-time warning.
19:15:03	krux02	like one that {.deprecated.} would provide?
19:15:17	*	kenran quit (Ping timeout: 265 seconds)
19:15:38	disruptek	like that, but without intimating that the procedure will be removed.
19:16:09	krux02	I think it should be removed.
19:16:35	krux02	And in fork of Nim it is already removed, and it won't come back.
19:17:49	FromDiscord	<Rika> then do you propose division be removed as it isnt exact for floats either
19:18:03	disruptek	huh?
19:18:19	FromDiscord	<Rika> wasnt that the argument
19:18:27	krux02	No
19:19:15	FromGitter	<kingramx> > *<disruptek>* @kingramx: hope it helps. ⏎ ⏎ It helped a lot ⏎ but I have another silly question: ⏎ Why retrieved data is surrounded by " " ? ... [https://gitter.im/nim-lang/Nim?at=5f4564334823780937418fd9]
19:19:43	krux02	Rika: The argument is: Decimal rounding for binary floating point numbers is actually no rounding at all.
19:21:12	FromDiscord	<Rika> why so
19:21:52	krux02	if you round 0.123 to 0.1 then it is the rounding that this function provides.
19:22:00	krux02	but actually it isn't
19:22:06	krux02	because 0.1 doesn't exist
19:22:29	krux02	and the result of the rounding will use the full internal mantissa to try to represent 0.1 as precise as possible.
19:23:08	krux02	In other words, you loose precision, yes, but you don't shorten the internal representation at all.
19:23:33	krux02	a rounded is supposed to require fewer byets to represent it, but the opposite is the case
19:24:06	disruptek	kingramx: surrounded by " "?
19:24:09	disruptek	~paste
19:24:10	disbot	paste: 11a frowned-upon behavior in chat; please use a service such as https://play.nim-lang.org/ or http://ix.io/ or https://gist.github.com/ and supply us a URL instead. -- disruptek
19:24:34	FromDiscord	<lqdev> @haxscramper that pragma… doesn't exist
19:24:44	FromDiscord	<lqdev> did you mean {.byref.}?
19:24:55	FromDiscord	<lqdev> it's used to annotate _object types_ tho and not variables
19:25:41	FromDiscord	<haxscramper> https://nim-lang.org/docs/decls.html
19:26:21	FromGitter	<kingramx> > *<disruptek>* @kingramx: surrounded by " "? ⏎ ⏎ yes, surrounded by " "
19:26:26	FromDiscord	<haxscramper> It worked beautifully until I tried to put in macro
19:27:03	disruptek	kingramx: what does that mean? whitespace you didn't expect? paste the code into the playground or ix.
19:27:24	FromDiscord	<Kaynato> A question about template/macro appropriateness for code substitution:
19:28:24	FromDiscord	<haxscramper> > did you mean {.byref.}?↵@lqdev Actually I get the same error with `{.byref.}`, which means the question now is: "I get 'invalid pragma error' each time I try to use pragma on variable in `quote do`'
19:28:39	FromDiscord	<lqdev> that's odd
19:28:51	FromDiscord	<lqdev> maybe `quote` is doing some funny business with idents
19:29:32	FromDiscord	<Kaynato> More or less, I'm thinking about something to take in `proc(...): T`, a list of args to pass into that proc to fill some preceding argument spots, and do some things with calling the proc with that arg and some local variables, and then returning a `seq[T]`
19:29:42	FromDiscord	<Lod> Is there any nimble tutorial?
19:30:56	FromDiscord	<Kaynato> Would something like `template findInfo*[T](fun: proc(varargs[auto]): T, args: varargs[auto]): seq[T]` be possible/appropriate? Or would a macro be suggested?
19:31:52	FromDiscord	<lqdev> `proc(varargs[auto])` isn't valid syntax
19:32:47	FromDiscord	<Kaynato> I'm guessing that there's no good way to pass-through-distribute varargs from template to function?
19:33:19	FromDiscord	<Kaynato> And/or that there are ways but that they are discouraged vs using macro in this sort of use case?
19:33:32	FromDiscord	<haxscramper> Yes, it was due to `quote do:`. Interpolating `byaddr` as variable into pragma solved the issue.
19:34:04	FromGitter	<kingramx> > *<disruptek>* @kingramx: what does that mean? whitespace you didn't expect? paste the code into the playground or ix. ⏎ ⏎ https://play.nim-lang.org/#ix=2ved
19:35:10	planetis[m]	krux02: to be clear, the math module has two round procs, one with a ``places`` argument, the other without, would you remove both?
19:36:50	planetis[m]	i have a use case for round(x: float): float but not for the other one
19:46:37	*	superbia joined #nim
19:52:35	*	quetzalb joined #nim
20:04:28	*	vsantana quit (Quit: leaving)
20:10:05	*	myphs quit (Ping timeout: 258 seconds)
20:13:29	*	natrys joined #nim
20:15:10	*	drewr` quit (Remote host closed the connection)
20:16:40	*	drewr` joined #nim
20:18:23	*	Vladar quit (Quit: Leaving)
20:18:54	*	drewr` quit (Client Quit)
20:19:43	*	drewr joined #nim
20:22:25	disruptek	!requires gram
20:22:26	disbot	gram: 11disruptek/ram 71 total
20:22:30	*	narimiran quit (Quit: leaving)
20:23:01	*	MarderIII quit (Ping timeout: 264 seconds)
20:23:18	disruptek	kingramx: https://play.nim-lang.org/#ix=2vev
20:24:00	disruptek	(you wanna getStr() on the JsonNode to retrieve its string value as opposed to merely rendering it)
20:26:27	Zevv	well PMunch, for what it's worth: https://github.com/zevv/nimfix
20:26:33	Zevv	it might or might not be usable, I dunno
20:29:25	*	apahl quit (Ping timeout: 240 seconds)
20:30:32	*	apahl joined #nim
20:31:15	*	awe00 joined #nim
20:32:25	*	awe00_ quit (Ping timeout: 240 seconds)
20:35:52	krux02	planetis[m]: Initially when I stumbled upon it, there was only the version with the two arguments (second argument with default value). I introduced the one argument round function, becaues it would make sense to have it and tagged the other one as deprecated.
20:36:09	krux02	so to answer your question, round with one argument is fine.
20:50:03	FromGitter	<iffy> How do I fix this concept/type error? https://play.nim-lang.org/#ix=2veB Error: invalid type: 'var AsyncSocket' in this context: 'StringSocket[asyncnet.AsyncSocket]' for var
20:52:59	FromDiscord	<dom96> You can't have an `Option[var T]`
20:53:11	FromDiscord	<dom96> You can't even have an `Option[var string]`
20:53:26	FromGitter	<kingramx> > *<disruptek>* (you wanna getStr() on the JsonNode to retrieve its string value as opposed to merely rendering it) ⏎ ⏎ TNX A LOT
20:53:33	FromDiscord	<dom96> `var` can only be used in the context of proc parameter types
20:55:14	FromDiscord	<Varriount> (for now)
20:58:44	FromDiscord	<dom96> @Varriount oh? How so?
20:59:32	FromDiscord	<Varriount> dom96: I believe @timotheecour had plans to do stack escape analysis
20:59:48	FromDiscord	<Varriount> To allow using `var` in structures
21:03:35	*	NimBot joined #nim
21:04:54	FromDiscord	<dom96> cool
21:04:54	FromDiscord	<dom96> With `ref` it's not a big deal not to have this feature though
21:04:55	*	ForumUpdaterBot quit (Read error: Connection reset by peer)
21:04:55	*	ForumUpdaterBot joined #nim
21:05:54	disruptek	🏀⚽
21:08:29	disruptek	leorize[m]: actions looks really good, thanks. found my first matrix snafu, too.
21:09:58	disruptek	we don't care about backporting arc fixes that are fixed in devel, right?
21:10:56	*	kenran joined #nim
21:11:23	*	thomasross joined #nim
21:15:55	*	kenran quit (Ping timeout: 240 seconds)
21:22:06	*	lritter quit (Quit: Leaving)
21:22:57	*	tane quit (Quit: Leaving)
21:23:51	*	Jesin quit (Quit: Leaving)
21:25:11	*	natrys quit (Quit: natrys)
21:26:04	*	Jesin joined #nim
21:33:47	*	Sembei joined #nim
21:51:51	FromGitter	<alehander92> wow dude
21:51:55	FromGitter	<alehander92> i like football
21:52:25	*	solitudesf quit (Ping timeout: 240 seconds)
21:52:26	disruptek	where's my password generator?
22:24:52	FromDiscord	<Kaynato> Is there a sane way to extract one arg type of a function referred to by a compile-time generated identifier in a macro?
22:25:47	disruptek	sure; you can get the type of the symbol if the macro takes a typed input.
22:27:07	FromDiscord	<Kaynato> The proc's symbol is generated at compile-time - would it be possible to, after generating the identifier, to pass it into another macro that takes in typed? Or, since those would all be NimNodes, would it be already wrong
22:27:20	disruptek	nah, it'll work fine.
22:28:00	disruptek	you can write procs that operate on nodes, too, of course.
22:28:07	disruptek	they will be compile-time procs.
22:37:10	FromDiscord	<Kaynato> Hmm, so trying to inspect the proc directly after `let procName = newIdentNode(prefix & suffix.strVal)` doesn't work, since at that point it's only an Ident and not an actual typed proc
22:37:40	FromDiscord	<Kaynato> But if I first return it from this macro, and then, through the template, pass it into another macro, then I can treat it as typed?
22:37:47	disruptek	sure.
22:38:33	FromDiscord	<Kaynato> There's no way to directly bind the ident in the local scope?
22:38:45	disruptek	there is, but it's not /in/ local scope.
22:38:53	disruptek	it's just data at this point.
22:39:13	FromDiscord	<Kaynato> Ahh, that tracks. So I have to dip it through the actual scope to actualize it.
22:39:28	disruptek	you have to dip it through semantic analysis.
22:39:50	disruptek	the compiler has to get hold of it and figure out what it means before it can answer questions about it.
22:40:27	FromDiscord	<Kaynato> Right. Node manipulation / AST work is done before imports?
22:40:51	disruptek	i dunno what that means.
22:42:22	FromDiscord	<Kaynato> Ahhh. Ok. So the problem itself is that I'm constructing a macro to wrap certain vk calls so I eventually get a seq[T] out of it.
22:42:24	disruptek	you can produce your own ast nodes that import. that should probably answer your question.
22:42:59	FromDiscord	<Kaynato> The type T is not defined by the vk proc's return type but is rather in `ptr T` as the last arg to the function, which itself I am generating the ident of in the macro
22:43:42	disruptek	you're not generating the proc, only the ident of it?
22:43:50	FromDiscord	<Kaynato> Yes.
22:44:04	FromDiscord	<Kaynato> I'm trying to sidestep the current clunky temp of just passing the desired return type into the template itself and have it infer the return type from the argtype of the proc to which the ident refers
22:44:35	disruptek	sounds doable.
22:45:05	*	awe00 quit (Ping timeout: 240 seconds)
22:45:29	disruptek	it might be easier to just use a compile-time lookup table.
22:46:55	FromDiscord	<Kaynato> Something like finding all the procs of that pattern from vulkan and generate a table mapping their names to their ostensible return types, right?
22:47:30	disruptek	yeah; consider that you can do this at compile-time and have a table with nodes, etc.
22:48:04	FromDiscord	<Kaynato> https://pastebin.com/raw/Y7GCuhL6
22:49:16	FromDiscord	<Kaynato> Currently broken pending edits to get rType from vkFuncName. I'm thinking that I have to first produce vkFuncName outside of vkGetMacro, get the lastArgType, and then pass that into vkGetMacro....?
22:49:34	disruptek	you've ruled out using the existing vulkan stuff made in nim?
22:49:42	disruptek	'cause there are a few...
22:49:55	FromDiscord	<Kaynato> I'm working off clybber's
22:50:09	FromDiscord	<Kaynato> Frag as far as I can tell is still WIP
22:51:23	disruptek	you're using untyped nodes which probably won't have what you need to actually compose the proc name.
22:51:29	disruptek	unless i'm missing something.
22:52:03	FromDiscord	<Kaynato> I've been using it like `vkEnumerate[VkPhysicalDevice](PhysicalDevices, state.instance)` with the previous pass-in-return-type scheme
22:52:48	FromDiscord	<Kaynato> Works perfectly fine - the template passes the nodes into the macro and generates the code
22:53:13	FromDiscord	<Kaynato> I could possibly compromise with something like `vkGet(vkEnumeratePhysicalDevices, state.instance)`?
22:53:39	FromDiscord	<Kaynato> That way I'd be passing the proc as-is directly to the macro. Hm.
22:54:11	disruptek	well, i'd want to get stuff typed as early as possible.
22:54:39	disruptek	it's hard to do much untyped. to put that another way, it's easier to write bugs.
22:55:15	FromDiscord	<Kaynato> First mode is probably just a little redundant. I'll probably settle for passing in the vk function as-is instead of cutting corners with composing names
22:57:19	*	awe00 joined #nim
22:57:26	FromDiscord	<Kaynato> Thank you!
22:58:21	disruptek	sure; i think you have a better sense of where you're going than i do. 😉
22:59:26	*	a_b_m quit (Read error: Connection reset by peer)
23:05:30	FromGitter	<alehander92> disruptek
23:05:40	FromGitter	<alehander92> dude i found my photos from hackathons in uni
23:05:46	FromGitter	<alehander92> i've had yellow-ish hair
23:05:57	FromGitter	<alehander92> we were all so naive
23:05:58	disruptek	lets see.
23:08:52	disruptek	at least your /went/ to uni.
23:09:09	disruptek	the only yellow hair i had was a girl named candy.
23:10:25	FromGitter	<alehander92> https://scontent-sof1-1.xx.fbcdn.net/v/t1.0-9/1003829_461438437310227_1085524231_n.jpg?_nc_cat=102&_nc_sid=e007fa&_nc_ohc=jbEY0PrJCKMAX9iNVrs&_nc_ht=scontent-sof1-1.xx&oh=5857183aa49e09703da176324c6e0936&oe=5F6A874D
23:10:38	disruptek	haha
23:10:47	disruptek	that's pretty bad.
23:11:10	disruptek	where's the hackathon code?
23:11:27	FromDiscord	<Clyybber> there is no code, it was all yellow hair
23:11:45	FromDiscord	<Clyybber> always has been
23:11:46	disruptek	oh, it was beautician's school.
23:11:49	*	kenran joined #nim
23:12:14	disruptek	yes, i realize 10 people are furiously paging through dictionaries now.
23:13:12	FromGitter	<alehander92> https://scontent-sof1-1.xx.fbcdn.net/v/t31.0-8/1404916_461722333948504_66767065_o.jpg?_nc_cat=105&_nc_sid=e007fa&_nc_ohc=kfXKWvLIKCIAX8p5U_0&_nc_ht=scontent-sof1-1.xx&oh=ed399fc86a2cbe173b1ba17b43d67180&oe=5F6BDEC2
23:13:17	FromGitter	<alehander92> here i am on a computer!
23:13:45	FromGitter	<alehander92> with probably free food
23:13:53	disruptek	those were the days.
23:14:02	FromGitter	<alehander92> in an incredibly technological env
23:14:10	FromGitter	<alehander92> i wanted
23:14:11	FromGitter	<alehander92> dude
23:14:19	FromGitter	<alehander92> i failed my team and my friend
23:14:30	FromGitter	<alehander92> i finally i wanted to combine music with auto code linting
23:14:43	disruptek	wut
23:14:54	disruptek	clyybber: where is your skiplists pr?
23:14:54	FromGitter	<alehander92> so while you read your code and you're on a bad section it starts some heavy dramatic music
23:14:57	FromGitter	<alehander92> iirc
23:15:16	disruptek	hmm, code smell as a service.
23:15:18	disruptek	it's not crazy.
23:15:37	disruptek	to quote duran duran, "i smell like a sound"
23:15:50	disruptek	so surely we can hear a smell.
23:15:56	disruptek	a code smell.
23:16:04	FromGitter	<alehander92> https://www.facebook.com/media/set/?set=ms.c.eJxFk8dtQDEMQzcK1Mv~_iwURP53jA0VZzVHaJmW~_MdX5E2AHBznaxkTq4xawkENK0~%3B1xtY3u0t97~_XI~_Hjt9Hvv5Tcl9rP3xyvnjcYCdnJfv1b9zui9YRcFFTrCRGyzk~%3BXvf9sunamDGK~_ZjQa7T2Y~%3BqXj359aMG~%3B9BvcXqwHkN8MN4R34x35C~%3Bm9~_svl~%3B2FQndyQWe~_vP1Ysr~_Ye8~_YP~%3BtY6a~_bVzbrqzz~%3Bm2cNmHrffrKabGDOrxPM9xrzNeq4h3nzGwcz38Cf9OMekvegi~%3Bj~_Z~_jczw70z2~_ix7V
23:16:04	FromGitter	... k5PenI571~%3Bh3IsZBvHjr0K~%3BahZMP8tMioR4N8~_9T~_~_jUXMHVHvDxuMN~%3B3~_0~_x9GP~%3Byv1bBJjxcfvSZD2Jfox6Xv~%3BB~_7RE~%3B0ou7Ef4XmHe0mTk5z1Yox9n~%3BX3~_fvX23WeUkm9~_kZzf3H569hfflO~_G.bps.a.461720503948687&type=1&__tn__=HH-R
23:16:09	FromGitter	<alehander92> oh no those links
23:16:11	FromGitter	<alehander92> ok i am stopping
23:16:16	Yardanico	just mirror it to imgur :)
23:16:26	FromGitter	<alehander92> sorry
23:16:30	Yardanico	that link looks like it'll take all my data
23:16:38	FromGitter	<alehander92> yeah i prefer my linters normal
23:16:40	FromGitter	<alehander92> and textual
23:16:46	*	kenran quit (Ping timeout: 258 seconds)
23:17:00	disruptek	why are you on gitter, anyway?
23:17:06	FromGitter	<alehander92> it's a bit eastern european
23:17:12	FromGitter	<alehander92> i cant pay my irc cloud service
23:17:14	disruptek	i can't find my can opener.
23:17:21	FromGitter	<alehander92> i know this is an absolute blast for normal people
23:17:29	disruptek	irc cloud service?
23:17:31	FromGitter	<alehander92> they would be like "paying .. your irc .. cloud .. dude "
23:17:42	disruptek	you can use my bouncer if you want.
23:17:46	FromGitter	<alehander92> but i am too lazy to setup my bouncer
23:18:05	FromGitter	<alehander92> and irccloud is just a very ok app and i was like ok give them money
23:18:21	FromGitter	<alehander92> how does this bouncer reuse work
23:18:31	disruptek	what?
23:18:57	FromGitter	<alehander92> "you can use my bouncer if you want"
23:19:01	disruptek	didn't i just open an old can of something yesterday?
23:19:38	disruptek	i can make you an account and then you can connect to it.
23:19:40	disruptek	what else?
23:19:56	FromGitter	<alehander92> ok, sounds .. good?
23:20:01	FromGitter	<alehander92> i have no idea how it all works
23:20:15	FromGitter	<alehander92> but it sounds legal
23:20:17	disruptek	lemme try to figure out my password.
23:20:30	disruptek	it's not the least legal thing i do.
23:21:10	disruptek	i just found a file named ruptek\u{17}u
23:21:44	disruptek	i think this is a bad sign.
23:22:13	disruptek	oh wait, do you have a static ip?
23:26:21	*	awe00 quit (Ping timeout: 256 seconds)
23:31:16	FromGitter	<alehander92> i
23:31:22	FromGitter	<alehander92> do i
23:31:40	FromGitter	<alehander92> how do i find out
23:32:06	disruptek	does your ip ever change?
23:32:22	FromGitter	<alehander92> i dont know man
23:32:25	disruptek	'cause if it does, i have to teach the bot how to let you in or something.
23:37:09	*	krux02_ joined #nim
23:39:25	*	krux02 quit (Ping timeout: 240 seconds)
23:41:26	disruptek	what's good about microwaves is they kill most of the bugs.
23:42:00	FromGitter	<alehander92> yeah i am trying to figure it out
23:42:05	FromGitter	<alehander92> but not quite sure
23:42:28	FromGitter	<alehander92> it's not a big thing right now
23:43:08	disruptek	good, 'cause it's sounding complicated.
23:43:30	FromGitter	<alehander92> :) no
23:46:11	*	ptdel joined #nim